How the discovery session works now
Discovery is adaptive, resumable, and designed to capture the right information rather than rush the user.
Understand what the AI is doing during discovery and why it asks what it asks.
Adaptive discovery
The AI does not just fill fields. It first builds the common project foundation, then adapts its follow-up questions based on the project type it detects.
- A new product project needs different follow-up questions from a churn or redesign project.
- The AI keeps asking until the answers are specific enough to be useful.
- Vague answers are treated as incomplete, not as confirmation.
What the system is looking for
Discovery should leave the system with a clear understanding of the product, the problem, the workflow today, and the conditions for success.
- Who the primary and secondary users are.
- What current workflow exists today and where it breaks.
- What the decision objective is for this project.
- What success metrics and constraints should shape the work.
Unlock behavior
Discovery no longer needs the user to manually confirm each micro-step. The system advances when enough grounded information exists.
- The next step unlocks when the current section is sufficiently grounded.
- If critical information is still missing, the AI keeps probing.
- The project does not move into research until the foundation is ready.