Chest v2 is an autonomous research agent. Tell it what you want to learn, and it plans the study, recruits and interviews your users, and surfaces the insights that should change your next decision. While you ship.
You approve and steer. It does the rest, and keeps going after the report ships.
Say what you want to learn. The agent returns a research plan to approve.
It sources and screens the right users, from your base or a vetted panel.
Real conversations by voice or text, probing deeper, adapting as it learns.
It synthesizes and pushes the findings that should change a decision.
Everything learned feeds a living map of your users, kept current over time.
The agent surfaces what should change a decision, ranked by confidence, with evidence and a recommended action.
An Atlas of user truths with confidence and freshness. When a finding ages, the agent flags it to re-validate.
Participants speak or type. The agent transcribes, probes adaptively, and transitions between themes naturally.
It sources and screens the right users (from your product base or a vetted panel) and tops up to your target.
JTBD framing, laddering, and adaptive probing: the rigor a senior researcher brings, encoded under the hood.
Account and session data stored in the EU, your own model keys so LLM processing runs on the provider you pick, AI disclosure to participants, and anonymization. Trust is the product.
We're onboarding a small group of product teams to Chest v2. Tell us where to reach you.