From Chat to Copilot: Hire AI as your product's PM

Every PM already chats with AI. This session is about going the extra step and building the workspace around that chat - so it can draw on your product context, apply your standards, and run reusable workflows instead of starting from zero each time.

We’ll do this in Cursor, by setting up a repository: a structured workspace with product context, rules for how the AI should work, and skills it can run on demand. No coding background required. Facilitator: Mikas Stonis

Duration: 2 hours

What we’ll do together:

- Familiarize ourselves with agentic terminology: what are agents, skills, rules etc.

- Build the repo scaffold (constitution, context, rules, skills folder)

- Create two working skills end-to-end

- Run both skills live and iterate on the output together

What you’ll leave with:

- A working PM copilot repo on your machine, which has your product context, your working context and can perform actions on your behalf

- Two skills you can actually use in your day-to-day

- A clear mental model for how to keep growing this into your own way of working with AI

Requirements:

- Install an agentic IDE beforehand:

- Cursor (recommended)

- Claude code

- Codex

Important: To participate in the workshop you will need a paid plan for the agentic IDE you choose (so either a claude license, chatgpt license or a cursor plan)

- Bring a couple of repetitive PM tasks or decisions you’d like your copilot to help with - they’ll become your first skills.