Do the refinement research before the refinement meeting
Refinement meetings exist to clarify requirements and surface blockers. An agent with codebase access does most of that research and questioning before the meeting, so the team only decides.
Written by
-
Benedikt Stemmildt Founder & Co-CEO
Published: 6 January 2026 Updated: 1 min read
Our refinement used to take 90 minutes per sprint. Now it takes 20.
The difference: we moved most of the work to before anyone sits down.
Refinement needs context, not a meeting
Refinement meetings exist to clarify requirements, surface blockers, and align the team.
But most of that work doesn’t need a meeting. It needs context. Here’s what we do instead.
Research the ticket before the meeting
We gave our PM an agent with codebase access. Before the meeting starts, it:
- Analyzes the ticket against existing code
- Identifies affected components
- Surfaces similar past implementations
- Flags potential blockers
PMs get faster technical feedback. Developers spend less time explaining architecture.
How to start: record your stakeholder meetings, feed the transcript to the agent, and let it draft the initial ticket with the technical context already included.
Ask the clarifying questions early
We configured our agent to review every requirement and ask questions like these:
- “What happens when no results match?”
- “How should this behave on mobile?”
- “Which existing component handles similar logic?”
- “What error states need handling?”
Questions that would surface in code review now surface before coding starts.
How to start: give your PM an agent configured with your architecture docs. An example prompt: “Review this ticket against [repo]. List every assumption the developer would need to clarify before starting.”
The first version was garbage
The first version was garbage. PMs ignored it. Developers complained it missed context.
We iterated for weeks before it became useful. What made the difference was feeding it the right architectural documentation.
What the meeting is left with
Refinement goes from 90 minutes to 20. The remaining time is for human decisions: priorities, trade-offs, and edge cases that need discussion. Everything else happens before anyone opens their calendar.
What’s breaking your refinement meetings: time explaining code, missing context, or unclear requirements?
Further reading
Written by
-
Benedikt Stemmildt · Founder & Co-CEO
More on Specification and review
- Do the sprint planning work before the planning meeting Most of what a planning meeting answers needs data, not a meeting. Agents split stories, map dependencies, order tasks, write evil user stories and check acceptance criteria before anyone sits down.
- Breaking the Loop: Why agents never finish on their own Agents in a software factory always find another problem, another ticket, another improvement. Only a second, much slower loop can say when the work is finished: the one that reaches the market.
- Fast agents push the bottleneck back to the requirements Delivery was built around slow coding. With code, review and deployment fast, questions return faster than anyone answers them, and the constraint moves to where requirements are written.
Scaling needs named owners
Coding agents scale across teams when named people own the platform, the team practice, the guardrails, and the measurement. Rollout & scale is where that ownership is built and handed to internal owners, in your stack.
Training on this topic: Agentic Engineering for Product Managers and Product OwnersWant to check one team first? The self-check: twelve yes or no questions, three minutes, instant result, no account. Take the self-check