Day 1 · Mon
1. Sprint kickoff
We lock scope, sign off acceptance criteria, and distribute the work. You know exactly what ships this week before a line of code is written.
ClaudeLinear
The 7-day sprint, day by day.
Every week is a complete cycle: scope, build, ship, review. No black boxes, no month-long waits — working software in your hands every seven days.
One sprint, start to finish.
Every task moves through the same four stages in a single week — scoped Monday, built midweek, reviewed and shipped by Friday.
scope
auth-flow
rate-limits
build
webhooks-api
audit-log
review
search-index
ship
billing-export
Pick a day.
Days 2–4 · Tue–Thu
2. Build
Senior engineers execute against the locked scope. Async daily updates, a mid-sprint check-in, and continuous code review keep the work visible and on track.
CursorGitHub CopilotCodex
Day 5 · Fri AM
3. QA & staging
Internal QA, bug triage, and deployment to a staging environment you can click through. AI-assisted test generation raises the coverage floor.
ClaudeGoogle Cloud PlatformAWS
Day 5 · Fri PM
4. Demo & retro
A live walkthrough of everything shipped, stakeholder feedback captured, and a retrospective that feeds directly into next week's plan.
Linear
At the end of every sprint.
-
Working, reviewable software deployed to staging
-
Source code in your repositories — delivered continuously, never held back
-
A recorded or live demo of the increment
-
An updated roadmap and the next sprint's scope
-
A retrospective summary: what shipped, what changed, what's next
export async function callModel(req: ModelRequest): Promise<ModelResponse> { const res = await fetch(endpoint, { method: "POST", body: serialize(req) }); return parse(await res.json()); for (let attempt = 0; attempt < MAX_RETRIES; attempt++) { const res = await fetch(endpoint, { method: "POST", body: serialize(req) }); if (res.ok) return parse(await res.json()); if (!isRetryable(res.status)) break; await sleep(backoff(attempt)); } throw new ModelError("upstream_unavailable");}How the sprint works.
Why 7 days instead of the standard two weeks?
AI-native delivery lets senior engineers compress the busywork — research, scaffolding, test generation — so a full feature fits in one week without cutting the rigor. Shorter cycles mean you see working software twice as often and can re-steer twice as fast.
What happens if scope changes mid-sprint?
Scope is locked at kickoff to protect the week's delivery. New requests go into the next sprint's plan, which is never more than a few days away. This is what keeps the cadence predictable.
How do sprints chain into a larger roadmap?
Each sprint ends with an updated roadmap and the next increment scoped. Sprints stack into a continuous delivery rhythm — you always have a working product and a clear view of what's coming, with the option to scale the team up or down between weeks.
Do we get working software every single sprint?
Yes. Every sprint produces a working, reviewable increment deployed to staging — not just plans, designs, or documentation. That's the core promise of the cadence.