Builders Camp

Here’s the easiest way to make your product squad AI-powered (even if nobody is technical): > pick a canvas where the whole team works together (I use Miro) > make sure AI runs inside it (critical) > map the rituals for each team member > write an AI context section for each team member > write an AI context section for each block > create a prompt library inside the canvas to structure inputs/outputs > bring screenshots, docs, decks, everything into canvas (more context) > if anyone wants to update the canvas, you need to ask AI to do it for you > ask for AI feedback on stuff everyone else is working on (a real copilot) Here’s what this will do for your team: 1. Build the AI muscle. You get used to using it every day 2. Give context to AI about what the team does. Gets better daily. 3. Creativity goes up, as you have more versions on the whiteboard. 4. You AI-outsource the heavy lifting. Everyone gets more time to think. 5. Your own AI gives you details about everyone’s work. You are more informed. 6. The whole team can context-switch faster, work on more fronts at the same time If your product squad isn’t working this way, you have no idea how much you’re losing. PS: if you don’t know what canvas to use, try Miro. Check here what made me choose them: miro.pxf.io/aO7mqM #MiroPartner 

If you’re building AI stuff AND you’re doing it with other people (eg: your squad), you can no longer do it outside a canvas without AI. Let me explain: I’m building an AI product. Think of a 16personalities quiz but for people who want to become product builders and don’t know what skills they still need to develop. In my case, if it wasn’t for Miro I would be completely lost. No joke. Why? [Btw, before answering here's why I picked Miro miro.pxf.io/aO7mqM ] Building with AI forces you to deal with way more stuff compared to a non-AI product. Here’s what I’ve had to do so far: 1. research logic/theory behind these assessments  2. map the categories and questions that matter  3. calculate the algorithms where each question contributes to scoring 4. create functions that call LLMs to analyse results (multiple models) 5. improve on the prompts, run evals, iterate on variables, change algorithms 6. write personal recommendations based on what we learned from you 7. connect to the bootcamps from my product school 8. create a personalised path, all correctly linked and tailored 9. build the whole UX, dynamic to what the AI returns 10. deal with security, backend, APIs, automations, logic This means I had to deal with: > UX / design > research > algorithms > prompts > evals > backend > debugging > automations > content > integrations > usability testing > security How do you do ALL OF THIS if you don’t have a massive canvas to work on? Impossible on a doc, slides or project management-like product. Some are technical, others are content, others are visual, others are logic, etc. Tens of mediums. And don’t forget the actual “project planning” ie sprint planning, roadmapping, syncing, retrospecting… How is this different from non-AI product? Results from LLMs are non-deterministic, and evals are not like QA. You need A LOT of eyes looking at the whole process. You know what’s also critical? AI next to all you’re working on, doing stuff with you, for you. Full context on the content, helping the team. Check the video, shows how I am doing it. Big learnings for all PMs that want their squad to be AI-powered: - build your operating system to work on top of AI.  - docs are no longer enough, you need many types of mediums at the same time - Work on something that has AI to support you on top of all types of documents (Miro for example) - Do stuff with other people at the same time. More eyes = better results #MiroPartner

This is just magical.