Mention OpenAI and ChatGPT springs to mind for most. That’s cute, but OpenAI’s toolkit extends far beyond that, offering numerous ways for businesses and employees to leverage AI.
Table of Contents
Method 1: One way is fine-tuning the API
This essentially means teaching ChatGPT to be an expert in a specific area that matters for what you do (e.g., summarizing text).
For instance, Indeed fine-tuned ChatGPT-3.5 Turbo to write personalized job recs to job applicants, which helped them boost their outreach from 1M to 20M messages/month. Impressive indeed.
Yesterday, OpenAI launched new features that make it easier for your developers to fine-tune GPT-3.5—deets here.
Method 2: Companies can also have OpenAI do the fine-tuning for them (dubbed assisted fine-tuning)
SK Telecom had OpenAI fine-tune GPT-4 to be proficient at having telecom-related convos in Korean—they saw a 35% uptick in performance as a result.
Method 3: Then there’s building custom models from scratch
This is for companies with treasure troves of proprietary data in a specific domain that they need ChatGPT to know, like law!
OpenAI built a custom model for Harvey (AI assistant for lawyers) that’s 83% more accurate than GPT-4.
Method 4: Finally, there’s ChatGPT Enterprise
This is buying ChatGPT-4 for your entire org, with the big perk being that your business data stays private.
What happens in ‘GPT stays in ‘GPT…
Other benefits of Enterprise include unlimited ChatGPT-4 access, 2x speed, and longer context windows—think of it as ChatGPT on steroids.
Just yesterday, OpenAI’s COO Brad Lightcap highlighted ChatGPT Enterprise’s “tremendous momentum,” with its user base jumping from 150,000 to 600,000 in the last 3 months.
A lot of companies have seen moderate productivity gains with Enterprise. For example, 92% of Holiday Extras’ staff are saving more than 2 hours weekly, using it for tasks like:
- Crafting content in multiple languages.
- Analyzing CSV uploads for data trends.
- Improving communication for senior engineers.
- Managing customer service requests.
- Expensive but worth it for big companies.
Expensive but worth it for big companies.