Iām actually gonna start in my copilot that I use as a product manager - Claude. The goal here is I have an end customer, Iām trying to build them a product, and I want to explore what itās like to use an AI prototyping tool. But first, I need to refine my requirements.
Iāve already onboarded Claude with context - thereās project knowledge, custom instructions, and lots of threads for different initiatives. Now Iām starting a new initiative as if Iām a product manager at Riverside, a tool for recording in extremely high quality for remote conversations.
The initiative: thereās a really big opportunity to improve our reputation with customers by highlighting when things go wrong, whether or not they have to do with our software. I give Claude context by dictating casually, like I would to my designer or eng partner sitting next to me.
After Claude gives me ideas, I say: āI like the idea of number three, a host-only alert system, but make it for both hosts and guests.ā Then I ask Claude to create a prompt I can take to Bolt - essentially a quick PRD that will be used for generating the prototype.
Why not just start in Bolt directly? Youāre gonna use AI as a thought partner for way more than just prototyping - thinking, planning, strategic decisions, brainstorming. Everything before and after prototyping is part of the product lifecycle. I keep context in Claude so I can maintain that thread and branch off to specialized tools as needed.
The prompt I give Bolt doesnāt include technical instructions intentionally. Iām letting Bolt make those decisions, just telling it what it should look like and how it should work.
In Bolt, itās pretty magical watching it make a plan, check things off its list, write code, and run a terminal. But when I click the demo buttons, nothing happens. It doesnāt know there was an error - Iām crossing my fingers hoping we get an error because thatās the best part.
When these tools detect errors properly, theyāll say āpotential problem detectedā and give you a nice button to attempt to fix. Thatās the kind of magic you see across prototyping tools.
ā”ļø Your AI copilot isnāt just for one task - itās your strategic thinking partner throughout the product lifecycle. Branch to specialized tools like Bolt for specific needs, but maintain context in your main copilot to connect all the dots.
