10 Comments
User's avatar
Vivek Singh's avatar

Great post, Danny and Ravi. I've been using Bolt/Loveable to build wireframes/prototypes for the engg. and it has been definitely helpful.

I'm curious about your thoughts on AI-first approach for PMs on Product Strategy, User Interviews, Prioritization and Roadmaps.

Expand full comment
Danny Martinez's avatar

Appreciated, Vivek!

Interesting question, I think some of those might warrant another post or two :)

High level thoughts, LLMs can act as a great “second brain” to support product when making decisions about what to do.

It can act as the assistant that never forgets a customer quote, bit of context etc and can catch “gaps” in the thought process stage.

Ultimately though, PMs will always need to “own” and be accountable for the final decision, AI can be useful in assisting (not delegating to).

Expand full comment
PB's avatar

Great post and examples. Having product designer does change the situation. Even then now PMs can contribute more to the user experience with vibe coding tools … and already our teams see the value

Expand full comment
Danny Martinez's avatar

Thanks, and I absolutely agree - a great product designer is still a must. That's something I've certainly taken away from using these tools...

Expand full comment
Muhammad Lawal's avatar

I think the part where the engineers get a the source code to build upon is underrated. I'm working on a project and haven prototyped on lovable, my engineer has 50% of the code to get started already. This certainly depends on the complexity of the project, but it's generally a great win.

Expand full comment
Sundar Subramanian's avatar

Great insights. I’m wondering if PMs writing these detailed prompts is the best idea - or collaborating with designers and engineers to do that as a team. Isn’t there a risk of the PM biasing the team with their idea of what good design looks like, or fostering over engineering by adding elements that might compromise the quality of the code base.

Expand full comment
Danny Martinez's avatar

Yeah, there is a risk of that.

In practice, I've found that it improves collaboration; you can whip up a prototype relatively quickly and share it with colleagues and customers for brutal feedback.

Re: code base, several times I've shared with an engineer, and they've told me the code is overengineered and they'll implement it differently, which is also completely okay.

Expand full comment
Dr Eduardo Calvillo's avatar

I agree, AI Tools help you to get more output of the time you dedicate to a feature

Expand full comment
Danny Martinez's avatar

Agreed Eduardo! - Any that you've found particularly useful?

Expand full comment
Dr Eduardo Calvillo's avatar

I use a lot ChatGPT to help me refine the job to be done, the user story, or create acceptance criteria.

TBH,I started with vibe coding a bit late, so I still don't use it much for prototyping, yet.

Expand full comment