You should be playing with GPTs at work

Lenny Rachitsky argues that product and business teams should start experimenting with custom GPTs now, since AI tools are often dismissed as a fad or too complex. He explains what GPTs are, how to build one by uploading context like roadmaps and brand guidelines, and why that context matters. He then shares 20 examples gathered from hundreds of responses on social media, showing uses ranging from copy writing and user research synthesis to goal setting, documentation, and lead scoring. The piece matters because it turns a vague hype cycle into concrete, low-cost workflows that PMs can test and adapt themselves.

01Key takeaways

• Pick one recurring task on your team and try building a custom GPT for it to learn by doing.
• Upload the same context you would give a new hire, such as roadmaps, values, and style guides, since output quality improves with better knowledge.
• Use GPTs to offload work teams dislike, such as documentation, survey summaries, or repetitive copy drafts.
• Create a shared GPT that holds requirements so engineers can query specs and brainstorm technical approaches.
• Watch data privacy constraints; limit uploads and use team or enterprise plans when you need to restrict access to company knowledge.

02Key sections

Why GPTs deserve attention now

The author points to macro forecasts and real productivity reports from coding and industry examples to argue that AI impact is already significant. He frames early AI products as looking like toys, a pattern that tends to precede larger shifts.

What GPTs are and how to build one

GPTs are tailored versions of ChatGPT that run in the cloud and can be shared. Building one involves describing the task in plain English, uploading knowledge documents, and enabling capabilities like web browsing or automation tools.

Use cases for product and design work

Examples include a UX writing GPT trained on brand voice, persona-based conversations, a GPT answering questions over past research, and a shared requirements GPT that engineers can query. These show how GPTs can reduce handoff friction across roles.

Use cases for operations, growth and customer work

Teams use GPTs to draft experiment variants, automate survey and sales-call synthesis, score leads, enrich signups, and produce documentation that people otherwise neglect. Many report large time savings and cost reductions.

Caveats and public resources

Limitations include missing product context when data privacy restricts uploads, and the need for team or enterprise plans to restrict access. The author closes with a short list of publicly available GPTs for product managers.

03From the post

“1. My favorite decision-making frameworks 2. Why you should add a work trial to your interview process 3. The ultimate guide to willingness-to-pay It’s become clear to me that everyone should be playing with GPTs (custom versions of ChatGPT) right now. It’s easy to think AI tools are a fad, too complicated, or unlikely to have real impact on your job. But you’re wrong. Consider these recent news items that caught my attention: 1. PwC forecasts that AI could contribute $15.7 trillion to the global economy by 2030 from increased productivity. 2. When Italy banned ChatGPT, the productivity of coders in the country fell by 50% before recovering. On the flip side, Duolingo reported a 25% increase in developer velocity when using GitHub Copilot, and Shopify has already written over a million lines of code with Copilot. 3. Nat Friedman (former CEO of GitHub) and Daniel Gross (former Y Combinator partner) are investing $100 million in a startup called Magic that’s looking to replace engineers. Prominent engineers are worried. 4. In the public markets, there’s a growing separation between companies…”

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.