Tools · Startup & strategy
Design Sprint
Describe a decision — get a tailored 5-day design-sprint agenda and every facilitation artifact.
How it works
Corpus-grounded facilitation (Knapp's Sprint + Brown's Rumble via the start-a-company corpus): a day-by-day agenda with timeboxes tailored to the decision and the time available, plus the artifacts a real sprint needs — How Might We reframings, Note-and-Vote ballots with a tally structure, the Four-Step Sketch worksheet (notes → ideas → crazy 8s → solution sketch), and a Rumble structured-debate script. Specific to the stated decision, cited to the canon — so anyone can run a credible sprint without a trained facilitator. (Distinct from the toolbox's insight-sprint-studio, which sequences an analytics study, not a product-design sprint.)
You bring
{ decision, team?, daysAvailable? }
You get
{ sprint_summary, days_planned, agenda[] (day · theme · goal · activities), how_might_we[], note_and_vote[], four_step_sketch, rumble_script, grounded_in, provenance }
Use it for
- →Big product/strategy call: describe the decision + team → a full 5-day agenda + every artifact, ready to run Monday
- →Generate just the converging tools: How Might We frames + Note-and-Vote ballots for a workshop you're already running
- →Hand the sprint's chosen direction to the customer-development engine (MF-175) for validation
See it work
example outputDecision: "Should our B2B SaaS move new customers to self-serve onboarding or keep white-glove?" — 6-person cross-functional team, 5 days available.
Design Sprint — Self-serve vs. white-glove onboarding
Planned for: 5 days · Team: 6 (Decider = VP Product; PM, designer, 2 CSMs, an engineer)
The decision is whether to shift new-customer onboarding to self-serve. This sprint maps the current activation path, sketches both models, and ends Friday with a tested prototype in front of five real new customers — so the call is made on evidence, not opinion.
Agenda
Day 1 — Map & Target. Goal: a shared map and one target moment.
- 9:00–10:30 Expert interviews + map — CSMs walk the real activation path; build the end-to-end map.
- 2:00–3:00 Ask the Experts → target — Decider circles the make-or-break moment to attack.
Day 2 — Sketch. Goal: competing solution sketches.
- 9:00–10:00 Lightning demos — borrow onboarding patterns from outside the category.
- 10:30–1:00 Four-Step Sketch — each person sketches a self-serve flow solo.
Day 3 — Decide. Goal: one storyboard to prototype.
- 9:00–11:00 Note-and-Vote — silent review, dot-vote, Decider's super-vote.
- 11:30–1:00 Storyboard — frame-by-frame the winning flow.
Day 4 — Prototype. Goal: a realistic façade. Build a clickable onboarding prototype; recruit five new-ish customers.
Day 5 — Test. Goal: five interviews, patterns named. Watch real users; decide self-serve / white-glove / hybrid.
How Might We
- How might we get a customer to first value without a human on the call? — opens the activation moment.
- How might we keep the "someone's got my back" feeling in a self-serve flow? — opens the trust risk.
Note-and-Vote ballot — Which onboarding model to prototype
- Pure self-serve — guided in-product, no scheduled call.
- Self-serve + safety net — self-serve with a one-click "get help" escalation.
- White-glove, lighter — keep the call but cut it to 20 minutes. Each member places 3 dots; the Decider gets one super-vote.
Four-Step Sketch
Notes (walk the map) → rough ideas → Crazy 8s (8 onboarding variations in 8 minutes) → a three-panel solution sketch.
Rumble — self-serve vs. white-glove
Ground rules: disagree with the idea, not the person; cite the map. Prompts: Where's our evidence each model activates faster? What would have to be true for self-serve to win? Close: Decider states the call and the one assumption the Friday test must break.
Grounded in: Knapp et al., Sprint; Brown, Dare to Lead (the Rumble).
Run it now
Plan a design sprint
Turn a big decision into a timeboxed sprint agenda — How-Might-We reframings, a Note-and-Vote, the Four-Step Sketch, and a Rumble debate script.
Prefer code? Call it over the API or hand it to your AI agent via MCP — POST /api/bicycle/design-sprint · plan_design_sprint. API & agent access →