The Automation Advantage (2021) provides a roadmap for building automation and AI in a modern organisation. From the different stages a business must go through on its automation journey, to the best ways to reassure employees worried about job destruction, The Automation Advantage shows leaders how to prosper in a future world.
Introduction: What’s in it for me? Automation will continue to change the world – Here’s some guidance for how you and your organization can succeed in this technological future.
Automating your systems and AI technologies should make business sense.
Business strategy is a key part of introducing successful automation.
Outlining your automation roadmap, and setting your business up for successful automation and AI integration.
For many organizations, there are obstacles to automating. To implement automation, leaders need to reassure and inspire their workforce.
In many organizations, there are obstacles to automating.
Successful automation and AI means constant innovation.
About the author
Table of Contents
Video and Podcast
Read an Excerpt/PDF Preview
Technology and the Future, Management and Leadership, Computer Science, Machine Theory, Artificial Intelligence, Robotics and Automation
The goals of automation are shifting. In the past, automation aimed to improve efficiency by replacing human workers with machines that performed tasks faster and more cheaply. In the new era of “intelligent automation,” machines use artificial intelligence (AI) to augment human decision-making, problem-solving, strategizing and creativity. Rather than focusing purely on productivity gains, today’s intelligent automation supports a business’s overall market strategy. This book presents a blueprint for instituting an automation system across your organization.
Introduction: What’s in it for me? Automation will continue to change the world – Here’s some guidance for how you and your organization can succeed in this technological future.
Over the coming decades, it’s looking likely that automation and AI will continue to transform the working world. And not just in the predictable ways – in manufacturing and heavy industry. Whole sectors, where these technologies are still in their infancy, from journalism to medical science, will be changed unrecognisably. If businesses fail to adapt with the times, they could be swept aside. Because, just as it revolutionized manufacturing, automation is reshaping other areas of work.
In these summaries, you’ll be given guidance and strategies for how to ensure that you and your business are not only able to adapt, but are also able to benefit from these technological advances. You’ll see things from the perspective of a leader, looking to integrate automation and AI in their organization. And, last but not least, you’ll learn how automation and AI can leverage the talents of human beings, rather than simply replacing them.
We’ll be taking you through this summary to The Automation Advantage by Baskhar Ghosh, Gayathri Pallail, & Rajendra Prassad. Though the content is geared toward a more intermediate or advanced listener/reader, you may be able to follow as a beginner to the world of automation and AI.
This is a rich, in-depth work with lots of valuable nuggets of wisdom, and so, in the interest of expediency, I’m going to walk you through the most powerful ideas that ultimately distinguish this title from others in its field. And so what you’re going to get here is a focus on the fundamental steps to developing your automation and AI integration, setting up a roadmap for automation maturity, and insights on why a human-centric approach to your automation strategy is essential.
In these summaries, you’ll learn
- what happened when an Italian newspaper employed a virtual assistant;
- what learning to ride a bike has to do with developing your automation strategy; and
- why a human-centric approach is crucial to ensuring a sustainable automation and AI integration.
Automating your systems and AI technologies should make business sense.
Let’s begin by defining a term which’ll come up frequently in these summaries: Intelligent automation. When the authors use the term, they mean: the use of machines to automate tasks like data analysis, and decision-making, including the follow-up actions and learnings that stem from those. Yes, these are typically tasks that are still performed by humans, which might set a few alarm bells ringing in people’s minds. But don’t waste that energy worrying – Ghosh, Pallail, & Prassad’s work shows us how we can make intelligent automation benefit teams and businesses in our ever-changing technological landscapes.
Take the example of one of Italy’s leading newspapers, Il Secolo XIX. Long at the forefront of innovation in the newspaper business, pioneering color printing, a digital presence and integrated newsrooms, Il Secolo XIX nevertheless faced serious challenges.
The leadership realised that if they didn’t change fast to keep up with the digital world, they’d fall behind. They knew that they needed to produce high-quality, cost-effective journalism to sustain and grow their readership.
Their solution? They introduced something called a virtual assistant. The AI technology would quickly check pieces of journalism for grammatical errors, while also scouring the internet for surprising links. This meant that journalists at Il Secolo XIX could reference sources and previous news items that might’ve escaped their attention – the end product being in-depth, expansive pieces of writing that readers would be hard-pressed to find elsewhere.
It meant that they could spend their time writing great journalism, while the machine did the tiring research and proof-reading. Rather than making the human writers at Il Secolo XIX redundant, the virtual assistant simply allowed them to do their jobs better.
This is a great example of how leadership saw an opportunity to not only improve their key metrics, like maintaining and growing their audience, but adapt so that the journalists could do their best work, at scale. It also shows us the necessity and value of aligning business strategies with intelligent automation and integrated AI. So, let’s get practical – how can you set your business on the right path by integrating intelligent automation with your business strategies and systems? Well, as a leader in your business, you have to begin by asking yourself the right questions: Where do you want to take your business or organization? What does the future look like for your industry? Will developments there affect how your business’ value is defined a few years down the line?” In answering these key questions, you can then implement a clear strategy that will allow your leadership teams to think through all the components and systems, resources and boundaries that need to be reviewed and aligned. Basically, what you need to remember here is that while the goal of automating repetitive work and freeing up human potential is a great one, it must also make business sense. For automation and AI integration to really take off at your organization, you need to have a serious strategy.
Business strategy is a key part of introducing successful automation.
So what needs to be done? Firstly, you need to understand how, through automation, your organisation intends to get a competitive edge; you need to be clear how it’ll give you that edge.
This might even involve a complete overhaul of your business model.
So, as you’re considering embarking on the journey, ask yourself questions like: “Will my organisation be bringing in automation or AI to compete on price, and focus on market-share gains? Or is the intention to out-innovate competitors, leave them in the dust technologically, with an eye on the long-term advantage that brings?”
Then, you need to narrow down which automation projects your business will focus on. Automation is a difficult and energy-intensive process, so you need to be sure that the right projects have been prioritized. If there is more than one automation project in development, ask yourself which of those presents the best business case? That’s one you should focus on. In addition, leadership in business and IT must ensure that they work closely when establishing the organization’s automation and AI integrations. Think of it this way: If your service-oriented business is taking a financial hit because customers loathe the experience of the service via your app, it won’t matter how sophisticated your tech architecture is, if it doesn’t address their needs. You have to identify what the pain points are for your customers, and use those findings to inform your next actions in tweaking the tech architecture. Once automation has been adopted, companies then need to keep improving it. It’s very tempting to think, right, we’ve automated one important area, now we can relax. But this is a serious mistake.
To remain competitive, businesses must treat each day as if it’s day one at a start-up. In terms of tech, they must continually innovate. The losses from missing out on tech innovation far outweigh those you might experience if you adopt new tech, which have not been thoroughly vetted. Yes, it’s risky to jump in head-first, but sitting out might mean missing out on the competitive edge of your business being seen as “current.”
Outlining your automation roadmap, and setting your business up for successful automation and AI integration.
Creating an automation roadmap will help you clearly identify how you can implement, track, and innovate your integration strategy. It should be developed on two levels simultaneously: One with a focus on business results, and the other on the journey to better automation maturity in your company. Let’s get into the business results roadmap now – We’ll cover the strategic model for a successful automation maturity process in a moment. Okay. The authors structured the business results process into three main phases. The first is titled, Establish. For this phase, you should explore and pinpoint where there’s potential for automation, across your business applications. The next phase is Scale. As the name implies, here’s where you develop, activate, and scale your solutions. Finally, in the third phase – Operate – you’re ready to widen your geographic coverage to better observe true value realization, and update your intelligent automation strategies based on your learnings. Remember this level of processes should be interwoven with your processes for ensuring successful automation maturity. So, as we delve into this part of your strategic roadmap, keep this front of mind. Okay; so now you’ve reached the point where you’ve aligned your business strategy with your automation and AI goals, what’s next? Integrating automation and AI into your organization. This is a methodical process, and to embark on that journey means passing through different developmental stages.
Think of them as like the different stages of learning to ride a bike. First, you must learn to walk. Next, perhaps, you learn to master a tricycle. Then, a bike with stabilizers. After taking these steps you’re likely to be able to finally ride a bicycle with confidence. The same can be said for a business looking to automate: you have to go through the paces – without that first step, you’re not likely to ever make real advances. You need to start at the foundational stages where you examine the current state of your infrastructure and gaps for improvement, before setting a sustainable and successful automation strategy that’s agile enough to grow with your business. These are the building blocks to successful automation and AI integration.
First is the tools driven stage. At this point, you’re focused on solving isolated, but recurring problems, specifically in the areas of implementation and where there’s room for improvement. This means that automation efforts here are generally fragmented, and siloed, with a limited overall value.
Next is the process driven stage. This is when closer, or external evaluation of current automation infrastructure reveals or exposes inefficiencies in processes that don’t generate value for the company or organization’s bottom line. Once these are examined, look for opportunities to optimize by paying close attention to the points of overlap and connection in your products and services.
Following the foundational and optimizing stages, your business should be ready to explore robotic process automation. Simply put, what you’re doing at this point is automating repetitive tasks, which generate quick wins, using software programs. For instance, looking at the customer service experience at a bank, from initial query to sophisticated answer – all of these can be provided by a chatbot.
The penultimate stage is the data driven one. With useful and reliable data, your business can benefit hugely from the insights they reveal. Solid data will allow you to be able to – for example – better predict product experience, which means you can shift your focus from the goal of reducing cost, to one where you’re looking at improving efficiency and being agile and responsive to the ever-changing needs and behaviors of your consumers or clients.
For instance, by running algorithms over different data sources, it’s possible to learn about a consumer’s purchasing behaviour. The business can then send personalized product offers to that person. So, pay a close mind to your data.
Finally, the top level of maturity in automation is intelligence driven automation. This is when computing power is set up to perform complex functions or tasks that have traditionally been limited to human beings. So, sensing or comprehending ambiguities, learning these, and then acting on them. To be clear, here’s an example: When AI in insurance companies in the US were able to detect patterns in claims management operations, it was able to flag fraudulent ones, which costs the industry billions, and directly impacts premiums in the average household, forcing them to pay between $400 and $700 per year for these false claims.
As I mentioned earlier, it’s important to start at the beginning — You have to learn to walk before you can even consider riding a bike. So, if you want your business to reach the final level of maturity in automation and benefit from advances in technology, don’t skip on all the necessary stages. Follow this guide and take your organization into a sustainable and profitable future.
For many organizations, there are obstacles to automating. To implement automation, leaders need to reassure and inspire their workforce.
Next, let’s dig into the practical implementation of intelligent automation in your business. Right off the top, it needs to be said: If a business fails to adapt with the times, they could be swept aside; and many will fail to adapt. Change is hard, and in addition to that, there are several reasons for this inability to adapt. Some of them are self-imposed obstacles, and others are due to external factors. But without going into too much detail about those, what the important takeaway here is, when faced with obstacles to technological adaptation, like skills deficits in your teams, resistance from leadership, or reliance on outdated tech infrastructure, the answer is to build an entirely new system with automation and AI at its heart. Now, let’s address the people factor. Automation in the information age, means leveraging people’s talents. That’s a key difference with the first wave of automation – where, the reality is, machines did take lots of labour-intensive jobs. Fundamentally, automation today means saving time by getting machines to do the things they do better, while allowing people to focus on their uniquely human talents – like creativity and innovation.
In this regard, what exactly can leaders do? The most important thing is to build support for new tech by showing everyone how it’ll help them do their jobs better.
When attempting to “sell” the benefits of automation, it’s important to avoid getting side-tracked explaining its technological features. While this might excite software developers, it does nothing to reassure everyone else.
The following example illustrates this point well: When the fossil fuel company, Shell, wanted to introduce intelligent, head-mounted displays for their miners, they had to explain why this was necessary. Rather than hyping all the computing wizardry behind their head-mounted technology, managers at Shell chose to instead focus on the practical benefits.
Miners do dangerous work – digging through the bowels of the earth is no joke. Real-time assistance can mean the difference between life and death. And so, to sell their new tech, company leaders at Shell described situations when a miner could use the voice-controlled device to send an image of a technical problem, and receive real-time assistance. In the end, the new tech didn’t need much in the way of a sales pitch – the AI was absolutely welcomed.
Another way you, as a leader, can demonstrate the benefits of automation, is through this simple exercise. It involves just a pen and paper. For this exercise, ask your staff about all the different aspects of their job, drawing smiley, frowning, or neutral faces depending on how they feel about them.
Then, ask how they’d feel if all the aspects of their job that made them frown were automated. Quite suddenly, automation doesn’t seem like such a bad proposition!
In many organizations, there are obstacles to automating.
Intelligence automation integrations should be a holistic endeavour. Everyone at every level of the company from business and IT leadership to the operational staff, should be engaged and active contributors to this project. But in order to do this successfully, Ghosh, Pallail, and Prasad suggest the following model which can guide your decision-making processes as you go. These are the four Ss that’ll help you reach the highest automation maturity level: Simple, Seamless, Scaled, and Sustained.
- Simple: This is the point where you’re performing a kind of diagnostic check. Take a close look at your IT and connected business infrastructures. Wherever you see deficits or room for improvement, retool your applications and architecture with the goal of simplifying your functional units or modularity.
- Seamless: You have to make sure that your automation and AI integrations connect harmoniously with your core systems. You also need to ensure that your company culture supports and embraces these integrations, and are also energized by, and invested in their successful implementation.
- Scaled: Once you’ve engaged your automation system and strategy, you have to put it into action and see how it holds up under real-world conditions that’ll test its agility, robustness, and sustainability. So, don’t just go in and do a hard switch. Automation integration should be done in steady, mindful steps, which take into account all the elements of your business’ strategy, culture, and goals. For instance, talent development is key when it comes to scaling. Like the tech itself, everyone in your organization should be prepared to be agile and have growth potential, as the landscape evolves.
- Sustained: Once you’ve integrated your intelligence automation, the work doesn’t end there. You should set up processes for keeping up with developments in your industry, recognizing emerging possibilities, and observing how other businesses are exploring these developments or trends. And again, thinking holistically about your automation integrations, make sure you maintain a company culture where everyone can share their automation challenges and solutions. Following this model, along with the other suggestions we covered, can transform your business’ performance, and ensure its strategic purposes are met.
Successful automation and AI means constant innovation.
In today’s world, companies can’t afford to sit still. Such is the pace of technological change that anyone caught dawdling will be left in the dust by their competitors.
Whether that’s traditional booksellers waking up to Amazon’s digital store, or taxicab firms watching Uber’s success, those that choose not to innovate will lose out. The same is true of those looking the other way as automation and AI begins to take off.
Innovation means not being afraid of being a pioneer. For some companies, when it comes to automation and AI, this is a difficult step to take. They believe that it’s stupid to jump in head-first when the tech is still so new. After all, what’s the harm of waiting for things to settle before leaping in? Yes, it’s true that the pace of change means that whatever systems are in place right now will become obsolete soon. However, this can make it seem like it’s never the right time to begin. All the while, organizations that sit on the sidelines will fall further and further behind.
While there might be regrets about investing in soon-to-be-obsolete systems, the real risk is missing out on all the learning dividends that come with adopting the technology today.
And once automation has been adopted, companies then need to keep improving it.
The truth is, the automation journey never ends.
You’ve just read our summary to The Automation Advantage, by Bhaskar Ghosh, Gayathri Pallail, and Rajendra Prasad. These are the main takeaways: You have to make a clear plan for building a robust automation and AI integration and diligently follow those steps. The purpose for your organization’s automation plan should align and work in concert with your business strategy. Fear of job-destroying automation is an enormous problem – and one that emphasizes the need to explain how new tech should augment human talent, not replace it. Remember, without a human-centric approach to automation, it simply won’t be sustainable.
Finally, the most important lesson for those looking to introduce automation and AI: never, ever, stop innovating.
About the author
Bhaskar Ghosh is chief strategy officer at Accenture, where Rajendra Prasad is global automation and intelligent assets lead and Gayathri Pallail is associate director for automation strategy and deployment.
Bhaskar Ghosh, PhD, is Accenture’s Chief Strategy Officer. In this role, he directs the company’s strategy and investments, including ventures and acquisitions, all offerings and assets, and Accenture Research. In addition, Bhaskar has management responsibility for Industry X and driving responsible business and sustainability services. He previously served as advisor to the CEO and group chief executive of Accenture Technology Services. Ghosh is a member of the Accenture Global Management Committee.
Rajendra Prasad is the Global Automation Lead at Accenture and heads a team that has helped organizations across the globe successfully implement and scale their intelligent automation transformations. He has spent over two decades innovating and defining frameworks for driving efficiency and managing change in software engineering. Prasad holds multiple patents and built Accenture myWizard®, an AI-powered intelligent automation platform.
Gayathri Pallail is the Associate Director for automation strategy and deployment at Accenture. She has implemented enterprise-wide automation-based solutions and successful change management to enable seamless adoption for over 500 clients across industries. A conference speaker, Pallail is an innovator in automation analytics, prediction models, and tools to deliver automation effectively.
Table of Contents
Foreword by Julie Sweet, CEO of Accenture
Preface: Sharing What Works
CHAPTER 1 The Intelligence Imperative
CHAPTER 2 Beware the Barriers
CHAPTER 3 Start with Strategic Intent
CHAPTER 4 Choose Your Spots and Map Your Journey
CHAPTER 5 Plan the Plan
CHAPTER 6 Architect for the Future
CHAPTER 7 Inspire the Transformation
CHAPTER 8 Sustain the Gains
CHAPTER 9 Relevance, Resilience, and Responsibility
From the global automation leaders at Accenture―the first-ever comprehensive blueprint for how to use and scale AI-powered intelligent automation in the enterprise to gain competitive advantage through faster speed to market, improved product quality, higher efficiency, and an elevated customer experience.
Many companies were already implementing limited levels of automation when the pandemic hit. But the need to rapidly change business processes and how organizations work resulted in the compression of a decade’s worth of digital transformation into a matter of months. Technology suddenly became the essential element for rapid organizational change and the creation of 360-degree value benefiting all stakeholders. Businesses are faced with the imperative to embrace that change or risk being left behind.
In The Automation Advantage, global enterprise technology and automation veterans Bhaskar Ghosh, Rajendra Prasad, and Gayathri Pallail give business leaders and managers the action plan they need to execute a strategic agenda that enables them to quickly and confidently scale their automation and AI initiatives. This practical and highly accessible implementation guide answers leaders’ burning questions, such as:
- How do I identify and prioritize automation opportunities?
- How do I assess my legacy systems and data issues?
- How do I derive full value out of my technology investments and automation efforts?
- How can I inspire my employees to embrace change and the new opportunities presented by automation?
The Automation Advantage goes beyond optimizing process to using AI to transform almost any business activity in any industry to make it faster, more streamlined, cost efficient, and customer-focused―vastly improving overall productivity and performance. Featuring case studies of successful automation solutions, this indispensable road map includes guiding principles for technology, governance, culture, and leadership change. It offers a human-centric approach to AI and automation that leads to sustainable transformation and measurable business results.
Video and Podcast
“This book is a practical guide for how to make an impact and realize value from automation. There is a strong focus on the real-world execution of an automation strategy that makes it an accessible must-read for both a business management and technology audience.” – Mike Crisafulli, SVP of Software Engineering at Comcast
“Consistently creating value with technology is hard. The Automation Advantage makes it simple. The authors lay out a comprehensive path using modern techniques that will be appreciated by both the thinker and the practitioner.” – Mark Spykerman, Chief Information Officer of AmerisourceBergen
“The pandemic has required businesses to hit the fast-forward button on adopting technology while becoming agile enough to be ready for what emerges next. The Automation Advantage couldn’t have come at a better time, providing a comprehensive road map to launch automation and AI initiatives at scale and accelerating technology-driven transformation.” – Paul Daugherty, Group Chief Executive of Accenture Technology, Chief Technology Officer of Accenture, and coauthor of Human + Machine
“A brilliant field manual for the coming Automation Age.”
—Richard D’Aveni, Bakala Professor of Strategy at Tuck School of Business at Dartmouth College and author of The Pan-Industrial Revolution
“As with any transformation effort, getting your people to embrace change can be challenging. The Automation Advantage clearly identifies common myths and barriers and how they can be addressed through a human-centric approach. This is NOT a book of strategy. It’s a book of execution that will enable your organization with actions to take in order to succeed.” – Andy Nallappan, Chief Technology Officer of Broadcom
“Successful intelligent automation requires you to get more than just the technology right. The Automation Advantage addresses this clearly and effectively by providing the guiding principles and steps for changing your organization’s technology, governance, culture, and even leadership style.” – Charlene Li, Senior Fellow at Altimeter and New York Times bestselling author of The Disruption Mindset
“The authors cut through the hype to explore why automation is the necessary discipline to ensure your processes provide the data—at speed—to achieve your business outcomes. The Automation Advantage takes us on a refreshing journey that aligns how enterprise operations leaders need to approach automation and AI in the virtual economy.” – Phil Fersht, CEO and Chief Analyst of HFS Research
“A great book, with ample examples of automation at work. It cements all the ideas and discussions I have had on this topic with the expert automation team at Accenture.” – Rajiv Kakar, Group Chief Information Officer of Thai Union
“A CIO needs vision, leadership, and the technical nous to execute successfully. A world with zero human touch application maintenance is absolutely attainable—a world with self-healing and self-configuring systems powered by AI. The Automation Advantage sets out a strategic approach to getting there, both from a technology and human perspective.” – Ed Alford, Chief Technology Officer of New Look
“We are entering a world of autonomous enterprises where analytics, automation, and AI converge. The Automation Advantage definitively shows how organizations can improve decision velocity and precision decisions.” – R “Ray” Wang, CEO of Constellation Research and author of Everybody Wants to Rule the World
Read an Excerpt/PDF Preview
by Julie Sweet, CEO of Accenture
The Covid-19 pandemic made it very clear that technology is a lifeline for economies, governments, organizations, and people. It not only helped us solve immediate challenges and stay connected with one another—it also changed the way we see and understand the world.
Automation, the focus of this insightful and extremely practical book, is no small part of this change. Like many other technologies, automation advanced greatly over the past decade with the rise of intelligent systems—encompassing capabilities like applied artificial intelligence (AI), industrial and process robotics, and service robots. Now, due to the impact of the pandemic, the pace of automation innovation is moving much faster than we anticipated.
There is no better guide to automation than Bhaskar Ghosh, Accenture’s chief strategy officer and former group chief executive of Accenture Technology Services. I have worked with him for many years, and his invaluable vision and counsel have helped countless companies through their digital transformations. In The Automation Advantage, Bhaskar and his coauthors, Rajendra Prasad and Gayathri Pallail, draw on their veteran expertise to explain how automation works—and how it works best. They cut through myths and misunderstandings to blaze a clear trail for any business leader looking to develop an automation strategy, create new value, and realize growth—all with a human-centric approach.
Because like all technologies, automation should never be considered in isolation—it’s always in the context of solving human problems. I often tell other CEOs that if someone comes to them and says, “I have an automation project” or “I have a blockchain initiative,” they should say, “No.” They shouldn’t even ask what it is, because no initiative should ever start with technology. Ultimately, technology enables; it’s about empowering people.
Currently, we estimate that we are at only 15 to 20 percent of what could be automated. We’re going to see that percentage rapidly increase, as automation continues to spread from manual processes, such as building cars in a factory, to enhancing mental processes, such as giving a consumer a consistently better customer experience. Automation is helping companies actively grow their businesses through speed, safety, quality, cost-efficiency, and resilience.
Seizing this opportunity, companies are investing in this technology and creating jobs. But are people ready and able to fill them? Are young people leaving high school with basic technology skills and digital literacy? The answer in many cases is no. Before the pandemic, we were not ready to address the growing global reskilling need that automation is bringing. Unfortunately, we still are not prepared. As companies, we have a huge responsibility to reskill, and it is imperative that we bring together government, educational institutions, and not-for-profits with equal speed to partner on addressing this need.
As we enter the postpandemic era, we are seeing a world awakened to an incredible opportunity to reimagine and rebuild responsibly and sustainably. Instead of seeing automation as a technology that competes with people for jobs, it can and should be considered a way to contribute to our shared success. By eliminating mundane and repetitive tasks, automation allows us to focus human attention where it’s most needed—on creativity, empathy, and critical thinking. The Automation Advantage helps us understand not only how automation can create value for a business but also how its unique combination of technology and human ingenuity can play a part in transforming the global economy into one that works for the benefit of all.
Sharing What Works
The year is 2019, and the place is a beautifully furnished meeting room in a modern office tower—the headquarters of a global company. We’ve come to town to meet with a management team considering a complex operational business issue, which we think presents a great opportunity for intelligent automation. We’re confident we can make a case for this with the information we’ve assembled, which anticipates the questions these executives will probably have.
First, they’ll likely ask: “What is intelligent automation?”
Answer: It’s the new era of automation in which machines are used to perform tasks formerly reserved for humans—tasks that involve analyzing data, making decisions, and learning from what follows once decisions are implemented.
Second, they’ll ask: “Why should we invest in automation like this—will it really make us a better company?”
Answer: Your competitive edge will still depend, more than ever, on your people’s talents, but you’ll be able to leverage and augment those talents to an unprecedented level.
Third: “Are other companies in our industry already moving in this direction—and is it paying off?”
Answer: Yes. (We’ve done our research.)
Finally, they’ll ask, “Given our other priorities, why is now the right time to invest?”
But as the meeting begins, the surprise is on us. These aren’t the questions the executives ask.
As we share our thoughts on known pain points in their type of business and how automation could help, the executives are already nodding and pushing us along. This is a group already convinced that the opportunities are exceptional. It turns out there have been pockets of experimentation going on in their organization, and these limited, one-off solutions have been yielding seriously good results. So this top team wants to take things to the next level.
They’re past the why questions and well into the how. As they aim to approach the automation opportunity at scale, they want to know where to start. How to identify top business cases and how to determine the priority order in which to address them. How to assess their legacy systems and data issues that might delay them. They also wonder how they can be sure their people—both the technology staff and the business users of the new tools—will value the change and want to be part of it.
It turned out to be the best boardroom conversation we’d had all year, as we found ourselves talking excitedly about the very issues consuming us personally at that moment. We’d been working with a variety of leading companies that had reached the same “scaling” point in their intelligent automation journeys, and we had been able to identify important patterns, the common pitfalls, and the clever solutions that cut across their companies.
As a technology consulting team with a practice area to lead, we had been pulling together these lessons and trying to distill them into practical guidance on designing and implementing successful automation projects. Now, the observations were jelling into insights that we knew could genuinely accelerate an organization’s progress.
And guess what: the following month, we had a very similar conversation with another company’s management team, also jumping ahead to the “how” questions—and then a great workshop with yet another right after that. We found ourselves repeating key messages and refining them, figuring out how best to organize and articulate the principles we were developing in practice. Eventually, we felt like we had covered all the major bases and developed a point of view on each of them, whether in a white paper, blog post, speech, methodology module, or presentation deck—however, we felt we should really pull them together into one coherent package. One of us named the challenge at that point: “That would require a whole book.” And here we are.
The book you hold in your hand is designed to answer that smart set of burning questions we continue to be asked by managers and leaders who want to proceed quickly and confidently in automation initiatives. But if we’re setting the whole context for how this book came about, we also have to mention that other big source of pressure that hit organizations all around the world in 2020—much as we might all want to forget it. We were well into the process of writing the book when Covid-19 hit, and we, our company, and all our clients found ourselves struggling to protect our people, help our customers, and remain productive amid a devastating pandemic.
We know and work with many academics, management gurus, and analysts focused on the realm of intelligent automation, and sometimes we envy their situations. For some of them, a lockdown on global travel and orders to work from home made this a year when they suddenly had more capacity to reflect and write. For us, it was a crazy time of demand for client solutions, as automation became more important than ever. This was an urgent realization by many firms.
As early as April 2020, Information Age reported that HFS Research surveyed 631 major enterprises about how they anticipated the pandemic would affect their strategies and operations. One of its questions asked: “How do you expect Covid-19 to impact your spending” in 10 major areas of technology investment. Fully 55 percent of respondents knew already that their spending on automation would increase, making this the second most reported increase after cybersecurity—both obvious concerns as so much mission-critical work moved abruptly to online platforms.1
For us, the challenges came simultaneously on many fronts—in our client projects, every one of which is unique, and also within our own company. Accenture itself moved instantly to new modes of staying productive—away from the consulting industry norm by which professionals spend the bulk of their time on-site at client locations and generate much of their value through in-person working sessions. The companies we had already been helping with automation seemed most eager to double down on that work. For companies seeking hyperautomation before (where discrete automated tasks roll up into seamlessly automated processes, and automated analysis identifies for itself new opportunities to take automation into new realms), the pandemic crisis pushed those ambitions further.
Having a book to complete in the midst of all this was not what we had anticipated when we signed the contract with our publisher. Yet there was a silver lining, too. We think the fact that we were so deeply in the mode of capturing and explaining principles from real-world engagements made it easier for us to make sense of this new, very dynamic, and frankly unnerving situation.
In many ways, this book represents the synthesis of long experience. The real veteran among us is Bhaskar Ghosh, who has spent a career as an automation advocate and been responsible for many groundbreaking solutions. He and “RP,” as all of Rajendra Prasad’s friends and colleagues call him, have worked together on many of these, ranging from small projects to large transformations. Since then, often with Gayathri Pallail leading the charge, we have helped many clients apply intelligent automation to remove pain points in risk management, sales strategy, customer service, and more.
As Accenture’s work in this area increased, we also found ways to make our own company’s internal processes more effective using automation solutions like workflow systems. The human factor is also critical. We’ve approached our internal automation opportunities with creativity, empathy, and innovation in the spirit of keeping the human advantage front and center for organizations we work with. In any situation where people must solve a complex problem, we’re convinced they can do it better if augmented by robust automation.
Along the way, we’ve seen our colleagues make many technical and managerial breakthroughs. In fact, as we began work on this book, the news came that Accenture had just been granted its 60th patent related to intelligent automation. To us, this is a real point of pride. We noted already that we stay close to the discoveries being made by leading researchers in universities, startups, and research and development labs. The work they do is invaluable to us. But meanwhile, our great privilege is that we are able to innovate in the setting of a global professional services company where we work with the world’s largest and most sophisticated companies. Their managers and leaders are inspirational. They are eager to get real innovation done. The problems they need to solve matter to their success and to the world.
Working where we do also means that we benefit from a culture that prizes discovery—and the documentation and dissemination of it to the world. This is evident from the constant stream of forward-looking research that Accenture publishes in every form imaginable—from podcasts to survey reports to high-level strategy books. In a company that invests heavily in staying at the cutting edge, it is a cultural value that we should have research-backed perspectives to offer our clients and other stakeholders.
But while we earn our patents for inventions that are specialized and technology focused, it’s critical that our client-serving teams can connect those with the problems of clients, even when the client involved is dealing with a highly strategic issue and talking in very big-picture terms. To our people, it may be clear that the patented tool or technique directly contributes to the client solution, but it isn’t always obvious to others. A book like this can equip more people in the business world with the story line that explains why this fast-evolving intelligent automation technology is so important to achieve business objectives, and what it takes to put it in place, at scale.
To be clear, this is not a strategy book—it’s an execution book. Our self-assigned mission was to produce a truly practical guide. So yes, we devote an opening chapter to offering some historical context on automation and making the case for why it is now the time to invest in cognitive technologies—but we then move quickly into content relevant to organizations who already sense there is a big opportunity and want to translate that enthusiasm into action.
We identify the most common barriers to acting on that sense in Chapter 2, as a way of setting up the value of what follows; all of it is geared toward overcoming these obstacles. In Chapter 3, we emphasize the need to be clear on strategic intent because a shared commitment to the business priorities is the necessary North Star to guide every aspect of planning and implementation. This strategic imperative segues directly into Chapter 4, which focuses on how to choose wisely from the seemingly endless opportunities most organizations face to automate some activity or another. Hint: it has as much to do with overall capability building and intelligent automation maturity as with the potential bottom-line impact of individual solutions.
Throughout, we place emphasis on what it takes to derive the highest value from investments, and we address very ground-level problems and considerations. In Chapter 5, we explain how to map out an implementation plan, put governance in place, and keep track of progress. In Chapter 6, we get into technology discussions on how to architect a future-proof automation—at a level that is understandable to not just the technology team, but also other functional executives and general managers. In Chapter 7, we discuss the “soft stuff” that all experienced leaders know is really the hard stuff: change management, reskilling, organizational culture, and other elements of people management. At a high level, we underscore the takeaway advice at the end of each chapter, but we urge you to delve deep into the stories that bring the advice to life and the research findings that back up its validity. Our objective in the text has been to write to that all-important midlevel that helps the most strategic thinkers in organizations understand the art of the possible and helps the most focused tool builders and technologists make the connections between their work and the overall strength and agility of the enterprise.
As you’ll see, the metaphor and concept of a road map turns out to be very important to that overall success. It gets introduced early and reiterated often. Chapter 8 emphasizes that this is not a journey that has an end; there is always effort required to sustain and build on the gains—and without that effort, there’s a real risk of backsliding. Finally, Chapter 9 gives a glimpse of where the journey may take organizations in the future, as they use automation to achieve new levels of relevance, responsibility, and resilience.
Our hope is that, with this book, we are putting a valuable map into your hands. It reflects the long road we have taken, as we learned from years of complex business challenges, and then the crisis of a global pandemic, as we helped our most sophisticated clients transform from manual operations to automated (and in some cases hyperautomated) ones. But now it is your map to keep drawing, as you embark on your organization’s own intelligent automation journey. We wish you every success on it. We are happy if our experience can provide some guideposts along the way—and we look forward to learning more from what you are able to achieve.
We are grateful to our colleagues for believing in the idea of capturing our automation experiences and sharing them with readers. We would like to thank all the experts and thought leaders who contributed to this book’s success.
We extend special thanks to Julie Sweet, CEO of Accenture, for supporting us as we wrote this book and for her vision and leadership in making a difference in this new world through intelligent automation. Her conviction in the advantage of automation has been a constant source of inspiration for us.
Our gratitude and sincere thanks to Accenture’s technology leader, Paul Daugherty, for believing in the value of this book and for providing his thought leadership, wisdom, and guidance throughout the process.
Deepest thanks to Julia Kirby who, with a brilliant editorial mind, played a crucial role in the book’s development. Julia dedicated many hours over the past two years to help us turn our ideas and thinking into outlines and finally into compelling chapters.
We owe special thanks to the many visionaries and extraordinary leaders at Accenture who took the time to read each draft and provide feedback to enrich the narrative. We are grateful to Kelly Bissell, Gregory Douglass, Kishore Durg, Edy Liongosari, Nirav Sampat, Rahul Varma, and Sanjeev Vohra. Thanks also to Francis Hintermann for not only reading the manuscript but also offering additional insights based on his extensive experience and research findings.
Throughout this journey, we were advised by Accenture marketing pros Kathleen Bellah and Raghavendra Rao, who helped us fine-tune our message. Kathleen generously read drafts and helped hone both strategic concepts and minute details. We are incredibly grateful to the rest of the marketing team, especially to Ed Maney for his dedication to this project throughout. Ed has worked with us week in, week out to make this book a reality. Likewise, Linda King has been an unfailing support to the project. Ed and Linda have been indispensable throughout the process from beginning to end, meticulously executing their work and providing guidance and encouragement. Special thanks also goes to Nancy Goldstein for bringing her marketing and communication expertise to bear to ensure the success of the book.
Huge thanks to the team at our publisher, McGraw Hill, led by Casey Ebro. Their support for the concept was unwavering, and they played a vital role in helping shape and refine the manuscript.
Our gratitude also extends to the many pioneering clients who have trusted Accenture to be their partner in their automation journey. We’ve had the unique privilege not only to research the ideas in this book but also to apply the concepts and observe the results as we worked with these true pioneers in this digital age.
We are blessed to have a brilliant team of automation leaders who have, day in and day out, implemented the concepts and ideas in this book with our many clients. Special thanks to Koushik Vijayaraghavan, Aditi Kulkarni, and Luke Higgins for the enthusiasm and passion with which they drive automation maturity for clients who have partnered with Accenture.
And finally, on a more personal note:
Bhaskar Ghosh: I’d like to thank my wife, Arpita, for all her support, encouragement, and inspiration for my work over the decades and for tolerating my long working hours on evenings, weekends, and sometimes also during vacations. Thanks also to my two sons, Anirban and Anindya, for their endless encouragement for this book.
Rajendra Prasad: I want to thank my wife, Kavitha, for her continued encouragement and support during my entire career. My two daughters, Janvi and Keerthana, help cheer me up every day with lots of fun. I also want to thank my mom, Saroja Devi, who first taught me how to write short articles during my school days.
Gayathri Pallail: I would like to thank my husband, Rajesh, for being unconditionally supportive at every step of my personal and professional journey. I’m especially thankful to my children, Ritvik and Anika, for putting up with me throughout this journey. Special thanks to my parents (Krishnadas and Jayanthi) and my parents-in-law (Ramadasan and Bhagyalakshmi) for all the support they have given me throughout my career.
Chapter 1: The Intelligence Imperative
A few years ago, one of Italy’s leading daily regional newspapers, Il Secolo XIX, introduced a new form of automation to its operations that, to some eyes, seemed audacious.
As a large, for-profit business serving a mass market, Il Secolo already depended on automation in many forms. The very genesis of its business was one of the most impactful automation technologies in history—the printing press—and more recently, it had transformed itself around the internet and digital technology. In fact, Il Secolo has long been one of the most forward-thinking papers in Italy, having pioneered color printing, integrated newsrooms, multichannel digital presence, and social media engagement.
This new automation project, however, struck some as moving in a dramatically different direction. This time, the tasks being automated were part of the intellectual work of journalists—the highly educated and creative workers at the very heart of Il Secolo’s premium product.
Like many newspapers, Il Secolo faced serious challenges. The number of readers was contracting and revenues were dropping. The leadership at Il Secolo XIX understood that sustaining and growing a loyal readership meant continuously rethinking what it means to be a newspaper. And doing so at the nonstop pace of a modern 24/7/365 newsroom. They realized they needed to find new ways to produce cost-effective, high-quality journalism to increase digital traffic, reader loyalty, and company revenues.
A virtual assistant that would boost its writers’ and editors’ productivity, without compromising quality. Consisting of artificial-intelligence-infused software, the virtual assistant was designed to leverage human talents and streamline the process of producing digital content.1
Now, when a journalist starts a story, the assistant continuously checks the text for data consistency, potential links to other sources, as well as spelling and syntax. By offering journalists prompts to other content it thinks will be relevant, whether from previous stories or external resources, the assistant gives journalists a completely new way to check sources, develop background understanding, and—more important—add extra content they might have otherwise missed.
Far from feeling threatened by their new, virtual colleague, workers in Il Secolo’s newsroom welcomed the support. In six months, every journalist on staff was using the technology. Many of Il Secolo’s journalists found the virtual assistant not only saved time but also stimulated new thoughts, prompting them to consider different angles, enriching their understanding, and revealing nonobvious connections to other stories already published or in the works.
At the enterprise level, the virtual assistant is translating strategy to reality. More abundant, high-quality content provides more opportunities to attract advertising and to grow revenues with digital subscriptions.2 Il Secolo’s success story is just one of many unfolding today under the broad banner of intelligent automation. It is a good case to start with because it has all the basic elements of the much wider trend.
Intelligent automation involves the application of smart machines, taking advantage of the various technologies that are collectively referred to as artificial intelligence (AI). It applies these tools to work performed by knowledge workers—a realm that has long been resistant to automation. And it calls for intelligent management of the automation process, to ensure uptake by the organization, to deploy the resources available for automation to the highest-value opportunities, and to integrate the new solution into a richly interconnected business system.
In Il Secolo’s case, automating a piece of the content creation workflow had implications for printing timelines, editorial assignments, ad sales, page layouts, and even human resources planning. All these elements had to be considered to create a coherent, structured plan.
Like many great newspapers around the world, Il Secolo regularly reports on advances in AI. One recent story, for example, highlighted software made by the American company Affectiva and the Japanese company Empath that is designed to detect people’s emotions. Perhaps aided by the virtual assistant, the article referenced the name of a famous science fiction story of the 1960s: Do Androids Dream of Electric Sheep? If it was possible to speculate in the 1960s, the reporter noted in the story, that computers were capable of unconscious cognitive activity, then maybe we shouldn’t be surprised that they could evolve to pick up on the unspoken feelings of others around them.
But even as reporters in this news organization have their eyes on the future of intelligent automation, they are, in their own jobs, making practical use of it very much in the present. They are part of a traditional business moving headlong into the digital age, reinventing itself to continue to thrive.
Automation’s New Era
Automation is not a very old word: it was coined less than 75 years ago, when a Ford engineering manager named Del Harder named the department that would oversee the company’s growing research and experimentation in replacing assembly-line workers with machines. Essentially, Harder took the noun automaton—meaning a robot or other type of self-operating mechanism—and turned it into the verb to automate. The term took off as fast as the technology. Automating work was a breakthrough productivity changer when it arrived in big industrial corporations in the mid-twentieth century.3
To be sure, the much earlier machines of the Industrial Revolution—starting with spinning jennies, power looms, water frames, and such—also represented automation, even if the term was not yet in use. But it was the scale attained by industrial businesses by the 1940s that turned automation into a discipline and constant quest.
Mass production facilities made the economics of designing, engineering, and deploying capital equipment attractive in a way they had never been before. As at Ford, the focus across industries was on making manufacturing processes faster and cheaper—as well as safer and more capable of producing consistent quality—by substituting machine power for human labor. Workers had always used tools, but automation meant handing work over to a tool itself.
By the turn of the millennium, machinery was pervasive across industrial settings, reducing to a fraction the former effort that had been required to produce goods. Automation had fully advanced into its next era.
If there is a single biggest story to be told about the transformation of the world’s leading economies over the twentieth century, it’s about their shift from being manufacturing-based to being service-based. Data from the US Bureau of Labor Statistics shows that, after a wartime high in 1943, employment in the nation’s manufacturing sector has dropped nearly constantly. As a share of total employment, manufacturing accounted for 37.9 percent of jobs in 1943. In 2018, it accounted for just 8.5 percent of jobs. Meanwhile, service jobs—first largely in retail and more recently in healthcare and social assistance—skyrocketed.
Today (and since the end of 2017), healthcare accounts for the largest number of jobs in the US economy, followed closely by the retail trade.4 With this shift came a growing awareness that the services economy, too, consisted of much repetitive, standardized labor—and a sense that automation could reduce the amount of human labor in services, just as it had done in manufacturing.
In banking, for example, Barclays Bank asked some 50 years ago: Could we automate the bulk of a teller’s work? And the ATM—the automated teller machine—was born. Within a decade, retailers spotted their own opportunity, and a grocery store in Troy, Ohio, made history in 1974 with the first barcode scan of a product (it was a pack of Wrigley’s Juicy Fruit gum).5
The questions have since become more ambitious. In healthcare, for example: Could automation make diagnosis of maladies quicker and more accurate? Could the ability to analyze trillions of data points across a patient population lead to more evidence-based treatments? Could surgeons be assisted in real time by virtual agents monitoring patient vital statistics and performing on-the-spot pathology? The healthcare system, with its many inefficiencies and an extreme need for best-practice performance, would seem to be ripe for much more automation—but only if the automation is intelligent.
This wave of automation was not limited, moreover, to those companies designated by government statisticians as “service providers” versus “manufacturers.” Even within manufacturing companies, because they had so thoroughly automated their production processes, payrolls were newly dominated by knowledge workers.6 Even at companies traditionally known as manufacturers, there are armies of people performing back-office and front-office roles, whose jobs consist mainly of information processing, greatly outnumbering the human beings on the shop floors.
We are now in the newest era of automation, when it’s clear that much of the office work that computers have enabled people to do more productively can actually be computerized. Whereas in the past, automation was relegated to the realm of largely manual labor, now it has entered the world of intellectual work.7 Again, it isn’t as though these managers, administrators, and professional workers lacked for tools before. As they performed their largely information-processing-oriented tasks, they made use of communications technologies, at least since the days of telegraphs and telephones, and information technologies as rudimentary as typewriters and adding machines. It was only when this toolkit evolved, however, into modern computing technology that it seemed possible that intellectual work might be automated in the way that so much of manual labor had been.
From Efficiency to Excellence
Automation projects were first embraced with the compelling, but limited goal of efficiency. Using a growing set of technologies, individual workers could be made more productive. The same goods in the same quantity could be produced in far less time, at far less expense, and with far less human effort.
At a time when product and labor markets were less complex, the direct effects of this efficiency on the economics of a business—the ability to offer lower prices to customers and the chance to expand profit margins—made the justification for investment in automation simple. Soon, however, as markets evolved and enterprises grew, another enormous benefit of automation became clear. Beyond boosting the efficiency of individual workers, investments in automation equipment allowed for more rapid scaling of enterprises into industrial powerhouses. (See Figure 1.1.) Factory automation equipment allowed for more tightly controlled quality, greater throughput, and supply chain optimization.
Now, in a twenty-first-century economy consisting of multinational corporations operating at global scale, managers are looking to an ever-evolving set of technology options to support the creation of business value that encompasses more than improved efficiency and can produce scale-driven gains.
We are seeing a shift from the era of industrialized automation to a new era of intelligent automation, in which the prior focus on the costs that can be cut transitions to a new focus on the customer experience, business excellence, service improvements, innovations, and smarter strategic decisions. In other words, automation increasingly is being viewed as a way to boost top-line performance as well as bottom-line savings.
FIGURE 1.1 Conventional to Maturity Automation Journey
As part of this, we are seeing fast growth in investments in automation applications not just for back-office productivity improvements behind the scenes, but also to support the front-office work of interacting with clients and customers. This is not an either/or situation; much value generation will still come in the form of cost reduction. But the emergence of intelligent automation will also reveal many more strategically important opportunities for top-line growth and enhanced service quality. Increasingly, companies will aim beyond cost-reduction targets and use automation to enhance the customer experience, the business’s ability to create value, and its revenue growth—which are all keys to remaining relevant in the future.
This new era of business value creation through automation will be marked by constant, rapid development of cognitive automation technologies. It will also be an era when managers expand their ambitions from implementing particular automation tools and solving one-off problems to establishing broad automation platforms that support and accelerate problem solving with automation throughout the business. We should expect the adoption of automation to grow exponentially as a result.
Indeed, this trend has already been established: investments in intelligent automation have increased dramatically in recent years. As reported in Forbes, one analyst organization predicts that the market for robotic process automation alone will add up to $12 billion by 2023.8
With so many companies now embarked on a transformational journey to become digital enterprises and learning fast from technology leaders, the foundations for more applications of automation in information-processing work are being laid. Combined with rapid advances in technologies related to AI, this digital revolution in business is creating endless possibilities for gaining value from machines that can sense, learn, and act.
Today, surveys find large percentages of organizations are already engaged in some form of intelligent automation. Many recognize it as a major technological breakthrough that has the potential not just to improve, but to transform the way they do business. Companies that aren’t experimenting with this capability risk falling seriously behind. They need to move quickly if they hope to remain competitive.
What Is Intelligent Automation?
As discussed previously, automation has been around for many decades. By now it has been applied across a large spectrum of business functions through all kinds of technologies designed to improve performance with human-machine combinations.
What makes an application of automation intelligent? In simplest terms, it only means that the automation solution relies on some kind of cognitive technology, such as (but not always) AI to arrive at good decisions or recommended actions. This is a good moment to underscore that while many people tend to use the terms AI and intelligent automation interchangeably, they are not the same. An intelligent automation solution does not always need the power of natural language processing, machine learning, neural networks, or other capabilities of AI—and certainly, many applications of AI have nothing to do with automating tasks in a production environment. The two intersect when the power of AI is utilized to take in historical data, find patterns in it, and make predictions based on it.
In their most sophisticated form, intelligent automation solutions can evolve their own capabilities to recognize problems and figure out how to solve them. Just as a human mind grows more knowledgeable and capable because of its inherent capabilities to associate cause with effect and learn from that feedback, machines equipped with AI can analyze data, make decisions, observe what follows based on those decisions, and make adjustments to try to do better. Each iteration serves to further refine the algorithm and increase the intelligence level of the automation. The machines learn by generating recommendations and self-remediate over time. In some situations, it becomes possible to automate automation—to use tools to spot opportunities for better use of tools. In most situations, learning systems can adapt and improve the more they work.
We already see the impact of intelligent automation in our lives every day. Netflix provides recommendations to viewers based on AI-powered personalization algorithms. Nike, the athletic footwear, apparel, and equipment company, has developed a system that customers can use to create designs for their shoes through an augmented reality experience and leave the store wearing them.9 Surgical robots that diagnose deadly diseases like cancer are demonstrating the impact of AI in healthcare. Across many industries, both business-to-business (B2B) and business-to-consumer (B2C), chatbots and virtual assistants are answering common customer service questions related to billing, product information, and services. Automation of these straightforward interactions proves popular with customers because it allows them to get things done at any hour, and far more quickly.
There are numerous ways in which intelligent automation is redefining possibilities and powering new levels of performance across business functions, from marketing and customer engagement to finance and accounting, and more. Smart tools—also known as intelligent systems—are being deployed to support optimized processes with sophisticated information-processing capabilities. Intelligent automation applies more and less advanced forms of automation to thoughtfully designed business processes to improve the performance of those processes—sometimes by orders of magnitude. For instance, intelligent automation can elevate the customer experience by not only managing predictable processes but also tackling more complex decisions to dramatically speed up systems and transactions.
Even without AI-level capabilities embedded in it, intelligent automation has the power to fundamentally change traditional ways of doing business, both at the operating level and at the level of individual workers and customers. These machines offer strengths (such as computational speed, accuracy, and the ability to cut through complexity) that are different from—but crucially complementary to—human skills. Rather than threatening those currently in the workplace, intelligent automation is invigorating workforces by changing the rules of what’s possible. People and technology are together doing things differently—and doing different things.
The Steps to Intelligent Automation Maturity
Beyond being a set of technologies and ingenious solutions, intelligent automation should also be thought of as an organizational capability. Gaining most value from it depends on growing an organization’s knowledge, skills, and other foundations.
Today, most companies have only nascent capabilities, and some have not even begun the journey toward intelligent automation. Only a small minority made investments very early and have devoted by now sufficient time and focus to grow what we would call “mature” capabilities. As always, however, the creative and experimental forays of these pioneering companies show others the possibilities and to some extent also pave pathways for them.
Our way of depicting the spectrum of organizations from least to most mature is to define five levels, reflective mainly of the why of their automation efforts—that is, what is driving them to invest in intelligent automation. What we have noticed is that as an enterprise matures in its capability, its managers’ understanding of the opportunity they are addressing evolves. Figure 1.2 depicts this evolution as rising levels of capability, and indicates the differences involved both in the technologies applied and the sophistication of the teams implementing and using them.
FIGURE 1.2 Automation Maturity Model
This is not to imply with these stair steps that a given organization’s journey has to be sequential—that a company cannot, for example, go to level four until it has gone through levels one through three. This is especially true given that companies are large, complex entities, in which different parts of the operation are often investing in automation solutions of different kinds. A back-office team, for example, might be using robotic process automation to relieve its people of routine, time-consuming tasks, while a front-office team is busy building a data-driven solution to take a service offering to the next level.
It is true that, to some extent, the five stages in the model represent a natural progression, whereby the accomplishments made at the lower levels lay the foundations for higher ones. Any automation solution integrating AI, for example, depends on the availability of data sufficient to train, test, and continuously operate the tool. Thus, a certain maturity with regard to data is a prerequisite to embarking on any AI-driven initiative. To a large extent, however, these levels are distinguished by a managerial mindset, as engagement with automation projects opens people’s eyes to larger opportunities to create meaningful business impacts. Let’s briefly describe what is going on at each of these five levels.
At the first level of automation maturity, we see organizations intrigued by exciting, tech-enabled ways to solve perennial pain points in their businesses and by learning what it takes to implement this new class of point solutions effectively. The focus is on specific tasks—how they are currently performed by individuals and how they could be accomplished better.
Automation efforts at this level are often fragmented—as experiments, pilots, and solutions are pursued in different corners of the organization by forward-thinking teams, making choices wholly independent of each other and the tools and approaches they will use. While the kernels of a capability have been planted in the enterprise, many of these solutions will yield benefits in excess of their costs.
The overall value reaped from automation, however, will be quite limited as long as these efforts take place in pockets and do not have the benefit of learning from each other. Still, these early successes inspire more efforts and allow the people involved in them to imagine larger-scale projects with more transformative power. In terms of both skills and mindset, this level lays the foundation for a new level of automation capability.
In the companies we would describe as having second-level automation maturity, individuals begin to realize that a point solution has a limited impact if other parts of the process in which it is situated are not also addressed. Often, a team that has succeeded in implementing a task-focused automation solution lacks the perspective—and often lacks the authority—to take on the larger process. Only when the team’s results are noted by others does the organization gain the will to take on the larger project of revisiting the process.
The narrow automation, in other words, serves to expose the inefficiency of the larger process. Of course, sometimes the results are worse than this: automation introduced without an understanding of a process’s complexities can cause more issues than it solves. Companies at this level of automation maturity know that tasks exist within larger processes, and they start by reexamining the whole process, often eliminating unnecessary steps. Usually, we see them using techniques like the Lean set of principles to streamline the processes.
The Lean process improvement methodology calls for the systematic elimination of activities that add little or no value to the business. It focuses on reducing time spent on non-value-adding activities and delivering products and services right the first time. It is a customer-centric approach, continuously evaluating whether products and services are delivered at the quality, cost, and speed that business users expect. When applied to a business process, it can transform the customer experience. The Lean management philosophy aligns well with Six Sigma methodologies for quality management, and with change management leading practices. Combining these disciplines gives companies a wide range of tools to measure, analyze, and improve processes.10
In the banking industry, for example, process-driven automation efforts have transformed the customer experience. Gone are the days of mountains of paper-based transactions—and customers being required to physically visit bank locations to accomplish routine tasks. Now, most leading banks equip their account holders to do their banking from anywhere and on the move. The banking applications allow users to make deposits, fund transactions, stop check payments, apply for credit cards, and perform many other activities like reviewing their account balances and credit card details, all remotely.
Robotic Process Automation (RPA) Driven
The companies we see as constituting a third level of maturity in automation are those that are broadly exploiting the power of RPA. Their focus is on automating repetitive tasks that can deliver quick automation wins, but they are setting up the infrastructure and learning processes that allow one project to learn the lessons of another, and a capability to be developed at a level above the individual project.
RPA is a vendor-supplied toolkit that makes it economical for teams within an organization to automate repetitive tasks that involve interfacing with information systems—despite the fact that the teams’ desire to automate this work does not rise to the level of being a priority of their internal IT organization. The robots involved are simply software programs that can easily be taught to perform a sequence of steps normally performed by an office worker to access, combine, process, and/or share information. Placing such routine tasks in the hands of the machines is not only a way to get them done faster and more accurately; the real value is that they free up the people to do less mind-numbing work.
RPA automates repetitive, rules-based processes that are predictable and involve high volumes of structured data. It emulates and integrates the actions of a human interacting within digital systems to execute a business process. Using RPA tools, a company can configure software to act autonomously, as a robot, to capture and interpret applications for processing a transaction, manipulating data, triggering responses, and communicating with other digital systems. RPA scenarios range from something as simple as generating an automatic response to an email to deploying thousands of bots, each programmed to automate jobs in an enterprise resource planning (ERP) system. An RPA software robot never sleeps and makes nearly zero mistakes.
As perhaps even this brief description makes clear, the processes that are typically automated with RPA solutions usually share several characteristics. They are routine, burdensome, high volume, and predominantly rules based. They have digital inputs, triggers, and few exceptions, and require only limited, predictable natural language interpretation. Where these conditions prevail, implementation usually is straightforward, and RPA quickly yields impressive benefits. It is not unusual to see cost reductions in the 50 to 80 percent range, as well as higher quality achieved through the avoidance of human error, and an 80 to 90 percent reduced time to perform tasks.
In many work settings, RPA-driven automation has already made everyday life easier. Thanks to RPA, paper forms are now digitized, data is inputted more quickly, claims are processed faster, and errors are rare. Popular applications include seamless onboarding of employees, streamlining the sales process by automating administrative tasks such as setting up a client in the billing system, processing most credit card applications, and more. Today’s customers interact with the organization across a range of touch points and channels, from chat and interactive voice response to apps and messaging. By integrating RPA with these channels, an organization can enable its customers to do more without live interaction with human representatives.
So this is not only an investment in efficiency—the primary focus of automation since the time of its inception—but also in effectiveness. RPA often produces very quick and measurable wins, so that once it is used in one area of an organization, it spreads rapidly to others. And therefore, a key step in the typical company’s automation journey is the point where management realizes that an organizational center should be established for these efforts that will build relevant knowledge of what works and save new projects from reinventing wheels. This stage lays the foundation for the journey to the higher maturity of organizations. It establishes and stabilizes a set of managerial activities that will be fundamental to pursuing data-driven and intelligent automation in bigger ways.
Businesses at the next major level of automation maturity exhibit much greater concern for data and how to manage it as a crucial enterprise asset. Based on what they have experienced so far, they are excited about the potential for automation to make many aspects of the business more agile and predictable, and they are acutely aware that good data is required to yield insights and intelligence. This phase lays the foundation for truly AI-driven intelligent automation and is where organizations can start to see its benefits.
Within Accenture, we have leveraged the power of data to transform many aspects of how we work. An easy example is in procurement. With offices worldwide employing more than 500,000 professionals, the company purchases a high volume of goods and services, amounting to nearly 200,000 purchase orders annually. For the global procurement organization, these amount to an average annual spend of several billion dollars, and the bills for these purchases arrive in the form of some 1.1 million invoices to be paid through the accounts payable function.
Managers realized that the processes involved in both procuring and paying could be much more optimized—in part because the individuals making buying decisions were often unfamiliar with the downstream accounting of their purchases. More generally, the processes were far from frictionless. There was ample room to take out inefficiencies and costs. In particular, the company applied intelligent automation to purchase requisitioning and non-purchase-order invoice processing by using predictive analytics and automating the recommendation of general ledger accounts to buyers at the point of purchase. Today, automated systems powered by predictive analytics equip buyers to be more accounting-savvy purchasers, and the effect downstream is a significantly streamlined accounts payable process, in terms of accuracy, time, and cost.
Most businesses today operate in more complex and dynamic environments than they did in the past. They need to balance the expectations of society, customers, and shareholders, and the tension between short-term competitive advantage and long-term sustainability. Data-driven automation can simultaneously address these challenges and help managers strike the right balance over time.
Far from being data-deprived, most organizations today are experiencing overflows of data from transactions, connected devices, and other sources. The computing power available now allows multiple algorithms to run over many disparate data sources. This means all the dots can be connected to reveal, for example, an individual customer’s spend pattern and typical purchasing behavior. That analysis would also highlight preferences for certain brands or product types and enable the sending of personalized product offers to customers. This is an important way in which many businesses are leveraging the power of data to become more and more personalized and productive. Abundant data opens opportunities to connect to customers on a hyperpersonal level, capturing their attention at the right time and place with the right message.
Data-driven automation can also generate a steady stream of insights to fuel intelligent technologies. It can make faster, smarter decisions to accelerate innovation. But if an organization has only highly fragmented or low-quality data, little can be done. That kind of data cannot be mobilized. Leaders need to reimagine their organizations’ data supply chains and processes to ensure transparency, trust, and accessibility. If high-quality data can be developed with all these characteristics, the return on technology and AI investments can be maximized.
Finally, at the highest level of automation maturity, we see organizations recognizing that they can and should implement intelligent automation at scale and across the organization. Already committed to managing data as a corporate asset and having already seen the benefits of automating information-processing tasks, these are the enterprises that are most likely to be adding AI to their automation agenda.
Today, everyone’s talking about AI and how it’s going to forever change the way we conduct business and live our lives. Many have commented on the idea of the “fourth industrial revolution,” speculating that the impact will be as great as, and potentially greater than, that of any prior technology-driven transformation. Whether or not it turns out to be the biggest technology revolution the world has ever seen, it will bring untold opportunities to reinvent individual businesses and drive revenue growth by augmenting the cognitive capabilities of human workers.
Start with the insurance industry, for example—an old and highly regulated industry. Insurance is still steeped in manual, paper-based processes that are slow and require human intervention. Even today, customers are faced with time-consuming paperwork and bureaucracy when getting a claim reimbursed or signing up for a new insurance policy. Customers may also end up paying more for insurance because policies are not tailored for their unique needs. In an age when most of our daily activities are online, digitized, and convenient, insurance is not always a happy customer experience. But today, a global push is underway by insurance companies to augment their technological capabilities so they can do business faster, cheaper, and more securely.
The past few years have seen heavy investment in AI by insurers, following decades of experience gained at lower steps of automation maturity. Most began at the level of implementing individual tools to perform targeted tasks more efficiently in areas such as claims management. Then they progressed to the level of process optimization, recognizing how advances in information technology allowed tasks to be eliminated and workflows to be accelerated. Next in their journey came robotic process automation, which allowed claims managers and practically any other kind of team engaged in routine information processing to hand over time-consuming task sequences to machines, freeing them up for matters requiring more judgment and creativity. At the same level, chatbots were introduced that could interact directly with customers and effectively respond to common requests—such as claims submissions or inquiries—via phone, email, or website chat feature.
At many insurers, efforts at these first three levels enforced a level of data management that made it possible to advance to data-driven, predictive solutions. In the claims management area, this has included automated analysis to discover fraud patterns and flag potentially fraudulent claims—an immensely valuable capability given that, according to the FBI, non-health-insurance fraud in the United States costs the industry $40 billion per year, causing the average household to pay $400 to $700 more per year in premiums.11
With so much automation of knowledge work accomplished already, insurers are well positioned to step up to the AI-driven solutions now possible for crafting more tailored and relevant insurance policies for individual customers and pricing them hypercompetitively. Only machines, with their vast data access and tireless processing power, can compile consumer-level offerings that customers recognize as including all the coverage, and only the coverage, they need—and at an attractive price. At this level of performance, not only do individual companies perform better, but the whole insurance industry prospers by appealing to a wider range of customers, including some who have never previously viewed insurance as worth the cost.
In many other industries, too, AI will take automation to astonishing new places, because it need not be limited to work that is strictly rote and rules based. It can be applied in areas that have traditionally required the human mind’s ability to resolve ambiguities, deal with exceptions and novel situations, and arrive at judgment calls that balance competing priorities. That opens a huge number of new opportunities for working with machines in new ways and redirecting human talent to more rewarding work.
AI enables a business to vastly improve how it interacts with customers. In some cases, this is thanks to the power of chatbots that converse with customers at any time of the day and can help deliver uniquely personalized and trusted recommendations that create more effective and relevant e-commerce or marketing experiences. The decisions of AI are model-driven, whether it’s an algorithm learning to play the highly complex board game Go better than a human; a use of computer vision to understand visual inputs with extraordinary accuracy; or predictive models that can forecast the future like never before. Machine learning and deep learning are at the heart of countless AI breakthroughs. AI enables a machine to continuously optimize its performance by learning from the success or failure of its actions.
Ethical Issues and Automation
As AI expands into areas of heightened sensitivity, such as human healthcare, it will be critical to subject the technology to greater human scrutiny.
We have seen only the beginning of the privacy intrusions, decision-making biases, and control concerns that can arise when work is performed autonomously by software and robots. Just as automation solutions are scaled up, so must be the management of ethics issues that come with them.
Any company that aims to have intelligent automation more widely applied across operations, more deeply embedded in customer solutions, and more responsible for decisions that affect lives—from medical diagnoses to government benefit payments, to mortgage approvals—must be deeply committed to the responsible automation principles presented in Chapter 9.
Too many AI applications today are effectively black boxes lacking the ability to explain the reasoning behind their decisions. As humans and machines work together even more, effective explanations will be at the very heart of this collaboration. The future of AI lies in enabling people to collaborate with machines to solve complex problems. Like any efficient collaboration, this requires good communication, trust, and understanding. Deploying AI now involves more than training it to perform a given task. It’s about “raising” it to act as a responsible representative of the business.
In their recent book, Human + Machine: Reimagining Work in the Age of AI, Paul Daugherty and Jim Wilson show that as humans and smart machines collaborate ever more closely, work processes become more fluid and adaptive, enabling companies to change them on the fly—or to completely reimagine them. As we look around, we see the journey to AI underway in every industry—and it’s picking up pace. The result of this rapid, broad-based adoption is that intelligent automation isn’t an option any longer. It is mandatory. The question is whether an organization has the capabilities to implement it across every aspect of its operations—and to reap the full benefits.
Fashion Designer Gives Wings to Designers’ Imaginations
One of the world’s fastest-growing fashion companies built an AI application capable of breaking down its product offerings into their various elements and then recombining those elements to suggest and design new concepts more heavily weighted toward attributes trending in popularity. This gives wings to designers’ imaginations, combining their creative art with the science of tracking trending products. Specifically, the development team leveraged AI techniques to help designers:
Analyze dress attributes and understand the market trends. The AI application is trained to identify the key elements of a dress and predict future consumer inclination from sales and margin data.
Create new apparel designs by recombining concepts from existing trending apparel. The algorithm takes the attributes of the most trending styles and generates new designs leveraging enhanced AI techniques such as deep neural networks.
Create digital variants with various colors. The AI algorithm generates different color variants of apparel using a style transfer approach.
Create digital variants by applying trending patterns. Trending patterns are identified using trendspotting algorithms and transferred to the apparel using the trained AI models.
The same concept can be used across industries. Expect to see such applications take hold in hospitality, home furnishing, advertising, fashion, and more.
Intelligent automation is a form of automation that brings higher performance to information-based work—not only by increasing its cost efficiency but by elevating the customer experience and boosting top-line growth.
Every company is currently at some stage of maturity regarding intelligent automation—and some are building that capability into a core competence, to be applied throughout the business.
Unlike past investments in automation, today’s must be “people first” in their orientation, designed to leverage human strengths and supported by investments in skills, experience, organization, and culture.
Chapter 2: Beware the Barriers
In most large organizations, the argument for investing more in intelligent automation is easy to make.
Intelligent automation can be applied to achieve higher levels of performance in numerous areas, spanning departments, geographies, and initiatives. Managers we talk with see possibilities for using it at every level of the business, from streamlining accounts payable to personalizing customer service to identifying acquisition opportunities. In a recent Accenture survey, 84 percent of business executives expressed the belief that their organizations would need to apply AI in their operations to achieve their growth objectives.1
In that case, why isn’t intelligent automation already more pervasive than it is? What’s holding companies back? In the same survey, over three-quarters of respondents reported facing barriers to automation and AI application. Fully 76 percent acknowledged struggling with how to scale intelligent automation across their business.2 For some, organizational structures got in the way, for others the worst problems had to do with data, and still others cited reluctance by employees to adopt the new tools. As with most capabilities promising the chance of great leaps forward, there are stumbling blocks.
A simple shorthand that management thinkers have long used to talk about the major elements that have to be considered in any substantial change initiative—people, process, technology, and strategy—applies equally well to examining where challenges most commonly arise in intelligent automation. This chapter looks within these categories to explore the biggest roadblocks getting in the way of some organizations’ journeys. It also, however, exposes some pseudo-barriers—that is, misconceptions and knowledge gaps that get in the way of success even though they shouldn’t. These are the myths of intelligent automation that cause hesitancy, false starts, or complacency. They are, in their own way, some of the biggest obstacles to be removed.
Barrier 1: A Shortage of Talent and Skills
Any company that intends to invest meaningfully in intelligent automation needs a diversely talented and reconfigured workforce to support and scale it. In survey after survey, business managers identify workforce issues as their greatest barrier: intelligent automation skills are in short supply, and therefore hard to find and expensive to hire.3
Across industries from insurance to education and across functional areas from IT to HR, hiring managers looking for developers, business analysts, program managers, and project managers are putting a premium on intelligent automation skill sets and driving salaries upward. Take the extreme case of a top-notch AI researcher. Peter Lee, a vice president inside Microsoft Research, once said that acquiring this level of talent is as challenging and expensive as acquiring a star quarterback in the National Football League.4
Short of that stratospheric level, companies are often more than willing to pay for talent, but they struggle to find it. In an O’Reilly AI Adoption in the Enterprise 2021 survey, respondents cited lack of skilled people and difficulty in hiring as the number one bottleneck to AI adoption, reinforcing the ongoing talent gap.5
Despite the shortage of AI talent, the commitment for retraining to develop AI skills internally does not appear as strong as what is needed. A 2018 Accenture study of 1,200 CEOs and top executives working with AI showed that even though almost half of business leaders identified skills shortages as a key workforce challenge, only 3 percent said their organization planned to increase investment in training programs significantly in the next three years.6
What kinds of skills are we talking about? Organizations need both the skills to develop AI and automation solutions and the skills to use them effectively. Clearly, a wide variety of talents go into the design, implementation, and scaling of intelligent automation solutions. Key competencies needed for automation engineers include skills in automation analysis, programming, software development, data analytics, data visualization, and IT security—on top of a firm grounding in ethics. In addition, organizations need expertise in the advanced technologies involved, such as robotic process automation, voice recognition, natural language processing, machine learning, and other forms of AI.
Moreover, at the team level, it can be a struggle to combine just the right mix of talent—in addition to the relevant technical skills. Teams should be interdisciplinary from the start, bringing together industry, business, design, and governance expertise in the right degrees. Some of these areas of knowledge might seem like nice-to-haves, but they all play crucial roles in giving a company an automation edge.
Meanwhile, on the receiving end of intelligent automation tools, there may also be important knowledge and skills deficits to address. Inevitably, automation affects job structures as it identifies the work that machines can handle and where human labor for oftentimes tedious tasks can be eliminated. Especially for workers in low- and middle-skilled roles, this may lead to significant job redesign, requiring some uplift of skills.
In some cases, the content of work will change to such an extent that typical education and training practices will not be sufficient. All automating organizations should be thinking, however, of how to develop the capabilities of the employees who will apply these new tools in their work, so they can both use them effectively and take advantage of freed-up time to add value in new ways. Successful implementation of sophisticated automation requires people to adjust to new ways of working.
Barrier 2: Organizational Resistance
If skills deficits are one side of the people barriers to intelligent automation, then the other side is made up of the organizational cultural challenges that can develop and hinder a company’s efforts. Often, initiatives run up against long-established habits, attitudes, and assumptions that make it hard for change to take hold, whether due to deliberate defiance or simple apathy—or just because everyone is running flat out with today’s workload and can’t spare the mental energy required to do things differently.
Low Cultural Flexibility
When Avanade conducted a recent survey about AI maturity, 80 percent of respondents agreed that business culture and change are the make-or-break items for AI’s long-term success.7
In a sense, change resistance is a larger barrier than any skills shortage. People who are eager to embrace a new way of working are often quick to learn whatever new approaches are needed, even without formal instruction. People who, on the other hand, lack any enthusiasm about a proposed change can manage to remain unaffected by even the most wonderful teaching methods.
Sometimes, as discussed in the following section, resistance is the result of workers feeling directly threatened by new technology. Other times, their lack of enthusiasm reflects skepticism about the value of the change, and a sense that precious time is being wasted on an experiment that will end up going nowhere. Perhaps most common of all, resistance is rooted in simple inertia. People have learned to get things done in a certain way, and as far as they are concerned, that process is working well enough as is.
As discussed more in Chapter 7, change resistance doesn’t always take the form of conscious pushback—and it should never be seen as insurmountable. None of the barriers discussed in this chapter are. But like all the others, it can derail an intelligent automation initiative if it is not recognized and responded to as a barrier. Just as in the case of skills deficits, change resistance has to be acknowledged where it exists and deliberately addressed.
Fears of Job Destruction
To be sure, there is plenty of uncertainty about intelligent automation in today’s workplaces, if not downright fear of what it means for the future of work. Millions of words have already been published about how the rise of AI and automation could destroy whole categories of jobs and leave even highly educated knowledge workers with no higher ground to move to. According to CIO Insight, 60 percent of people surveyed specifically about the prospect of their organizations applying intelligent automation believe workers will lose their jobs as a result. No wonder that the same survey finds 72 percent of respondents at the C-suite level saying the adoption of advanced technologies has been limited by employee resistance and unreadiness.8
It’s true that intelligent automation has the potential to seriously disrupt labor, and it is already doing just that. Some traditional jobs will become obsolete. Seeing this, however, as a straightforward transfer of labor from humans to machines is a vast oversimplification. There’s very little factual evidence to suggest that mass unemployment or widespread redundancy of human workforces will result from growing automation. In fact, it is just as possible that a more productive economy, brought about by the increased efficiency and reduction of waste that automation promises, will provide workers with more attractive options for engaging in value-generating and well-compensated pursuits.
Employers are generally looking at cognitive technologies as a means of augmenting the highest-value strengths of their human workforces and enabling their people to work more safely, more creatively, and more empathically. This opens up opportunities for businesses to leverage employee talents for more strategic and transformational programs. Successful enterprises have always tagged and aligned their automation transformation programs with the business transformation agendas so as to accelerate the creation of value.
Barrier 3: Subpar Processes and Outdated Policies
Process problems can appear in two fundamental ways in intelligent automation efforts. First, the processes and policies being subjected to automation can be suboptimal, poorly defined and implemented, or interconnected with other business processes in ways that are not well understood. Second, the project management processes being used by the automation team itself can be flawed.
Suboptimal Work Processes
In the first category, the biggest of these problems is the surprise that confronts many teams as they embark on what would seem to be straightforward automation projects. As they get down to the work of identifying the parts of a workflow that could be handed off from people to machines, they map out all its steps at a detailed level. That’s when it often comes to light that the process was not well designed at the start. It might have been optimized at some point in the past, but not revisited as steps within it changed or new technologies were introduced. More likely, it was never optimized, but was simply a way of doing things that people tried and, once satisfied, adopted it as a standard operating procedure. People continued to follow those steps and, perhaps with some workarounds and exception management, got good enough results. But it is far from the most efficient or foolproof process to accomplish its goal.
Veteran managers all know this time-honored principle: don’t automate a bad process, or you’ll only succeed at taking the wrong steps faster. Sometimes it’s best to begin with a clean slate and redesign the process in light of current reality.
Let’s say some long-established process calls for three approvals. Is that really necessary, or would two approvals suffice? Sometimes it’s helpful to look at a process from the perspective of its customer rather than its contributors. Are their attempts to interface with it more time-consuming or convoluted than they need to be? Could some steps be combined to streamline the process and improve the customer experience?
The point is process doesn’t have to be a barrier to success in intelligent automation if an organization pauses to get it right before automating parts of its operations.
Sometimes the barrier that trips up an automation solution the worst is never considered by the team because it isn’t strictly in the scope of their creative problem-solving work: it’s a corporate policy or set of policies that as written throws up a brick wall to implementing an essential part of the new approach.
Experienced automation leaders know that their project plans must explicitly include a timely review of corporate policies related to the solution design. And they understand and anticipate that time and effort may have to go into modifying policies that conflict with the project. Typically, this is not as politically difficult as it may sound, because many policies are easily proven to be antiquated or designed to address old issues or old processes that no longer apply. On the other hand, many legal, financial, and human resource policies are in place for very good reasons and remain immovable. Coming to grips and adjusting for those realities early on can be critical to an automation team’s success.
Poorly Defined and Implemented Processes
Attempts to automate an underlying process can also run into complications if the automation approach is opaque. This is the case when a process is basically sound, but there is little accurate documentation or transparency around it. It is difficult to proceed without knowing all the steps—the tasks, sequences, and actions that the process comprises. As we think about it, it strikes us this is why intelligent automation has taken hold so quickly in certain areas of IT, operations, production, and finance. They follow well-known, easily mapped sequences.
Take an IT process like password management, for example, or logging of service requests. The steps are unambiguous. The same goes for a process like maintenance scheduling in operations, or the tasks involved in invoicing and accounts payable in a typical finance department. Compared to these clearly specified processes, many other business processes feature more ambiguity and variety. Before they can be appropriately automated, they need to be similarly well defined and understood.
Similarly, the fact that various processes in a business interact in complex ways can present a barrier to intelligent automation. Few processes exist in such isolation that they can be carved out, retooled substantially with new technologies, and then expected to still interface seamlessly with upstream and downstream workflows. Managers hoping to apply automation to their groups’ work may need to negotiate with other process owners less interested in making adjustments.
Even where a process can be redesigned on its own for greater efficiency, it may not produce much impact if other processes are left as they are. Consider an insurance company applying intelligent automation to speed up the work of drafting an individual customer’s policy. That might not have any effect on the customer experience if the separate process of customer approval is not also accelerated.
The Process of Automation Itself
All these barriers have to do with the process that is the target of automation, but it’s important to note that automating work is a process in itself that can be more or less well mapped out and managed. When it comes to how companies go about developing and rolling out their intelligent automation solutions, most are held back to some extent by outdated processes for managing automation projects. This, of course, is a major topic of this book and the main barrier we and our colleagues help clients to overcome. Chapter 4 returns to the subject.
Barrier 4: A Technology Environment Built for a Previous Era
Technology barriers can seem like the least of an organization’s worries with regard to intelligent automation. With machine learning and other AI tools rapidly advancing, sensors proliferating, and computing power growing steadily, new hardware and software tools are finding their way into practical applications across virtually every sector of the global economy. Huge and growing numbers of solutions are already available from vendors, and many of them are producing impressive results. Yet most companies encounter barriers as they try to make use of these technologies. The barriers show up in the form of legacy architectures, inadequate data, and off-the-shelf solutions that are not nearly as turnkey as their vendors can make them seem.
First, there is the enormous challenge of innovating with legacy systems. The typical organization’s IT stack—spanning software applications, data, hardware, telecommunications, facilities, and data centers—was built for an earlier age, when none of the people laying those foundations could envision a cloud-oriented world of analytics, sensors, mobile computing, AI, the Internet of Things (IoT), and billions and billions of devices. That is the world that exists today—and we should fully expect a new wave of revolutionary changes to reshape it into a different world of tomorrow.
A legacy architecture constitutes a barrier when it makes it hard to bring in new applications or make enhancements to applications built on that old foundation. It is important to note that most companies’ enterprise architectures were not only based on different technological foundations; they also were built on a presumption that the functions they supported would be stable and enduring. They did not anticipate the constant introduction of new applications that is the reality of most companies today.
Every new application raises the same question: What dependencies exist between the application being introduced and other applications also on that architecture? There might be 25 other applications that must be taken into consideration before work can begin in earnest on the one new solution that was envisioned. Say, for example, that an online retailer becomes aware of a new payment option and wants to make it available to its customers—perhaps a hot new mobile app is gaining in popularity, and a retailer risks losing sales if it cannot include that new option in its ordering process. To add that new payment option on the legacy architecture, the retailer would have to first assess what impact it would have elsewhere in that architecture.
This is not a new problem; the barrier presented by legacy systems has always meant that there is a high bar for a new solution to be considered worth doing. It grows successively worse, however, every time a company adds a new application. Each successive generation of additions brings more complexity as old interconnections are disrupted and new ones are created.
By now, the conventional IT stack has truly reached its practical limit. Until a business goes to the very substantial effort of redefining both its business and IT architectures, it will have to deal with barriers as it tries to integrate the advanced tools of AI. Indeed, given the revolutionary potential of AI, these new architectures should be AI-centric—constructed at every layer in ways that place top priority on supporting it.
As we enter the 2020s, many companies’ IT leaders would describe the barrier they face as a lack of a microservices architecture. This very modern style of architecture enables greater flexibility by decoupling data, infrastructure, and applications—three elements that in traditional software solutions are bundled together. Think of this as an architecture consisting of boxes—small groupings of applications that are as independent as possible, so that making a change to one of them will have minimal ramifications downstream or elsewhere in the system. Now, picture the opposite: hundreds of applications with their links impossibly tangled together, their interconnections having been complicated by generation after generation of applications. That is the picture of today’s barriers in most organizations; it presents a nearly impenetrable thicket against new automation applications.
Another foundational barrier for most companies is the state of their data. This is a barrier particular to intelligent automation because especially where machine-learning algorithms are deployed, the systems require data in prodigious volumes.
Few companies complain that they have too little data; today’s digital mechanisms for information gathering and processing produce a tsunami of data points. The problem is that too little of it is usable, either because of poor quality or because of limited accessibility. Often there are difficulties in integrating data stored in different formats. This is exacerbated when mergers and acquisitions result in separate pools of data that should be combined but aren’t. One recent study estimated that 97 percent of a typical company’s decisions are made using data that its own managers think is of unacceptable quality.9
The way to think about this barrier productively is to see it not as a lack of adequate data but as a lack of good data management strategy. Usually, the data exists in some form for companies to produce real-time insights and actions, but it remains locked in unstructured and semistructured forms. It has to be cleansed, engineered, and optimized to be useful to decision-making and automated action. One bright spot is that automation itself can help. So-called capture software, for example, now exists to automate data capture in currently unfriendly forms and to do the content analysis, identification of key elements, and extraction required to make it useful for an enterprise’s business systems and applications.
Disappointment with Turnkey Solutions
The number of vendor solutions available in the realm of intelligent automation has exploded in recent years. Just look at the startup space. Entrepreneurial activity and funding of new companies is so high that the annual list produced by CB Insights of just the fraction of AI startups that are “most promising” includes a hundred new companies each year.10
It might seem like the only barrier is an embarrassment of riches from which to choose. But it’s important to understand that these solutions do not have the generic level of functionality that IT departments usually get from off-the-shelf enterprise software. In particular, to the extent that the software needs to be trained on a specific company’s data, it is not reasonable to expect a plug-and-play solution from a vendor. A lot of work still has to go into making it work in the company’s own environment.
The fact that vendor solutions typically don’t work right out of the box presents a barrier that sometimes surprises company managers eager to make progress with intelligent automation. Worse, it may cause them to gravitate toward only those small-scale point solutions that can be used without much tailoring or training. In that case, they may face different kinds of barriers: not being able to scale the solutions and failing to think holistically about what would help the business achieve its most strategic goals.
Barrier 5: Lack of Strategic Alignment
When intelligent automation is managed most strategically, the thinking happens on two levels. First, there is a clear plan for building an intelligent automation capability within the organization, stating its major goals in clear terms and outlining an action plan well designed to achieve them. Second, there is alignment of the intelligent automation solutions pursued with the overall business strategy. Writing in MIT Sloan Management Review, David Kiron and Michael Schrage refer to this pair of strategic perspectives as having a “strategy for and with AI.”11 Insufficient attention to either of these levels is a barrier to success.
A common strategic barrier in intelligent automation efforts arises where managers have not taken the time to specify their goals in objectively measurable terms, based on current baseline performance numbers. If a specific project aims, for example, to enhance the performance of a call center, it is not enough to state a goal like “we will automate the process of opening case numbers.” The project needs to specify, for example, that it aims to reduce handling errors by 40 percent and cut processing time by 80 percent. Implied here is that the team would have an accurate awareness of today’s handling error rates and average processing time. These become the key metrics by which the project’s progress (and ultimate success) will be gauged.
A lack of salient metrics becomes a barrier to future automation because it leaves intelligent automation proponents unable to point to proven benefits. But it also presents barriers to the effective management of the current automation project, as the team lacks the clear targets it needs to keep its efforts on track.
At the outset of an automation effort, a team should spell out the end value it hopes to realize, in whatever forms that value will take. Cost savings are often targeted most prominently, in part because they are easier to measure and can be used to show a project’s return on investment. But intelligent automation can also be applied to increase speed or enhance quality, perhaps by simplifying a process or by giving time back to customers. Often, improvements along these two lines have far more impact than incremental cost savings to a company’s competitive advantage, growth, and long-term success. It is vital not only to establish clear metrics, but also to measure what really matters.
No Road Map or Strategic Plan
Along with clear goals, a strategy for automation needs to spell out how the team will achieve those goals over a planned time horizon. A major barrier to high-impact success in most companies is the lack of such a road map. Many have not developed this disciplined way of managing initiatives because their adoption and implementation of automation, analytics, and AI has so far been done in piecemeal fashion—in a spirit of experimentation and in pockets of the organization that happened to have high interest and sufficient skill. Without a structured road map, however, they cannot progress as fast or as effectively as they could.
Any project needs its own road map, but at the highest level, a company’s overall road map should represent its resolve to realize the full potential of intelligent automation. It should map the journey from its current automation starting point to an automation destination at which business goals are routinely being set and met across the business by employees augmented in creative ways by transformative automation technologies.
As Chapter 4 discusses much more, an intelligent automation road map is essentially the strategy for how automation will be developed and introduced for greatest cumulative impact. Companies that lack this road map are held back by a patchwork of ways of working and fail to take on the challenge of updating systems that were built for the business needs of years ago. The older technologies they depend on are not up to the challenge of meeting tomorrow’s business goals.
No Sound Basis for Prioritizing Projects
We just discussed how intelligent automation should be approached as a structured journey, with a “you are here” road map to simplify this step and help teams determine how to proceed. That journey, moreover, is not just a journey of enterprise technology deployment. It is a voyage of transformation at the level of the business model and certainly the operating model. As Kiron and Schrage insist, having a “strategy for AI is not enough. Creating strategy with AI matters as much—or even more—in terms of exploring and exploiting strategic opportunity.”
The worst strategic blunder is the one that has plagued IT departments since they were first established: a lack of alignment between their projects and the priorities of the business. Without such alignment, intelligent automation projects often lack a clear business case. A surprising proportion of organizations are hobbled right out of the gate because they have not made a compelling case for how intelligent automation will lead to competitive advantage in the marketplace.12 Thus, they lack a defensible position on precisely what they should automate and in what order, and struggle to persuade leadership to commit to an intelligent automation program at meaningful scale.
Making strategic choices requires having a way to prioritize, but a big barrier for many companies is the lack of an opportunity assessment process. This results in a situation that might be described as “a thousand initiatives blooming.” We recently visited a multinational business that had more than 300 separate automation initiatives in play—yet none of them were connected. Leaders were approaching automation on an ad hoc basis, adopting tools to solve problems identified by particular teams, and for the most part picking low-hanging fruit. Some technology-driven organizations with hundreds of technology options have many trials happening simultaneously. Other enterprises take a fragmented approach to applying automation—identifying opportunities à la carte and implementing discrete tools to solve isolated problems.
We see some companies automating prolifically—but targeting the wrong things. Should the priority be a global operations and supply chain solution? Perhaps an inventory management process? A fix for training and development gaps? It is possible for intelligent automation to add tangible value to almost any process or decision by dialing in speed, efficiency, problem solving, and adaptability. Yet enterprise capacity for change is limited in any given time frame. Strategic choices have to be made. A company must have a process of sorting the opportunities into categories such as quick hits, strategically important advantages, and so forth.
To be sure, there can be good reasons for starting small, proceeding with caution, and launching experiments. It is typical for a company to simply look for repetitive tasks to automate, on the assumption that the financial benefit will materialize. But if efforts are too siloed, the synergies that could be exploited enterprisewide are being lost. Pursuing automation in disconnected silos limits the power and enterprisewide efficiency that automation could deliver. A lot of companies get stalled in pilot phase or early-stage AI adoption. The problem is that they are implementing the technology in silos and discrete projects. This approach not only lacks an enterprise-level view, but it also introduces increasing layers of complexity and thwarts scalability.
At worst, it means that resources are being spent on automation efforts that don’t ultimately benefit the business much at all. Take the example of deploying a virtual agent to automate some standard reporting. Automation of reporting is a very common wish-list item for managers because it seems obvious: reporting is tedious work and so rule-based that it is absolutely conducive to automation. And yet, what is the impact for the business to have that activity automated?