RIP PRD
PRDs serve a singular purpose: communicating what needs to be built with alignment across stakeholders. I have written some detailed PRDs in my career; AI changes this.
A picture says a thousand words, and I have always believed that a good wireframe/detailed mockup was better than tens of user stories. With generative AI and codegen, the cost of building prototypes is nearly zero. It is accessible to everyone, and if the product/engineering org is set up well, anyone should be able to prototype within the Company’s design system. In such a world, detailed PRDs that define the what and the why for communicating across stakeholders are no longer needed. Product Managers should be able to go from a prioritized feature/product request to a working prototype in hours. A product document that captures the various assumptions, tradeoffs, and GTM ideas will still be helpful, but the PRD will be much more concise.
In fact, if you follow the Amazon method of starting with a press release before building a product, your PRD is a lightweight index file to:
- project background
- DRI/RACI details
- link to linear project
- link to Slack channel
- links to prototype(s) and/or wireframes/mockups
- link to GTM docs
- tracking plan
- log of open questions and decisions
Maybe a more appropriate title for this post was “RIP User Stories” or “RIP JTBD”?
Maybe a PRD should be written retrospectively to capture the various decisions/tradeoffs before launching the product?
There is an argument that extremely detailed PRD can help create a valuable context for AI codegen. I am not yet convinced.
(Credit: @n2parko)