ROSC Interview Tool

Team 4766 · Haddad-Ball Medical

Ask your team lead for the code. It's stored only in this browser tab.

ROSC Interview Tool

Team 4766 · Haddad-Ball Medical — capture interviews, track how the model changes, build the report-out
Please don't include patient-identifiable information. On submit this is organized into notes + insights and saved to the team's record.

Living Business Model Canvas

Every hypothesis the team is testing, colored by the honest evidence in saved interviews: supported · contradicted · mixed/neutral · dashed = untested. Hover a hypothesis for its evidence. Add or remove hypotheses here; new interviews score against the canvas as it stands. New interviews score themselves against the whole canvas automatically; after editing the canvas, use Apply interviews to re-score the stored interviews against it.

Look up a hypothesis

Type a hypothesis — or any part of one — to see every interview that returned a verdict on it, split into supporting and contradicting. Searches the canvas menu and hypotheses that only ever appeared on an interview. Click any interview to open it.

Proposed by interviews

New assumptions the engine spotted in an interview. They stay off the canvas until you accept one. Edit the wording first if it needs tightening.

What the interviews taught

Quote-verified learnings from every interview, including ones that validated no hypothesis. Anything under (no block yet) is a pain your canvas has no hypothesis for: that is where the next hypotheses come from.

Hypothesis-change synthesis for the weekly Team Report-out

Generated by the model from every stored interview — what each hypothesis was, what we learned, direction, and why. Everything below is copy-ready for slides.

Who to interview next candidates only — nothing is contacted

Searches the web for real people who can speak to the segments the record is thinnest on, and proposes them with a source link and a Mom-Test opening question. Every candidate is unverified until you check the source yourself. This never emails, messages or adds anyone — it only produces a list.
Runs on your team's own server · self-contained (no Airtable) · powered by your Anthropic API token