Handbook / The engagement
How a Cohorte engagement runs.
Walking through the 12-week Team Bootcamp end to end: the structure, a sample week day-by-day, the mentor cadence, the sponsor cadence, the deliverables, the capstone. The shape that scales across cohorts; the exercises inside each week are customised per stack.
The 12-week structure
Three phases. Twelve weeks. One shipped system at the end. The same shape across every Team Bootcamp; what varies is which exercises run against which stack, and which mentor is matched to the cohort.
| Week 01 | Kickoff & LUMEN | Sponsor kickoff. Participants nominate the project they will ship. LUMEN scoping methodology introduced. Each participant writes their first brief. |
| Week 02 | Verification | Self-consistency, conformal bounds, exploitation-surface analysis taught against the participants' real projects. |
| Week 03 | Governance | Governance log design. AI Act / ISO 42001 mapping for the use cases on the table. |
| Week 04 | Phase-1 close | Briefs locked. Verification designs sketched. Week-4 sponsor report written and delivered. |
| Week 05–08 | Applied build | Participants build against your real data on your infrastructure. Weekly mentor 1:1. Live session digs into the technical layer the cohort needs that week. |
| Week 09 | Pre-capstone | Verification design hardened. Operating brief drafted. Mid-phase sponsor pulse. |
| Week 10–11 | Ship | Production gates wired. Governance logs connected. Final sponsor briefing prepared. |
| Week 12 | Capstone | Capstone demo open to sponsor leadership. Final sponsor report. Engine Room community access opens for all graduates. |
A sample week, day by day
What week one actually looks like for a participant. Live sessions are recorded. Mentor 1:1 is scheduled around the participant's availability. Applied work between sessions is real (on your data, not on a sandbox).
Sponsor kickoff (60 min, sponsor + cohort) + cohort intro (60 min, cohort only)
Sponsor opens with the business outcome the engagement is aiming at. Cohorte founder lays out the twelve weeks, the cadence, what success looks like by week 12. Cohort introduces themselves and their candidate projects. By end of day, everyone knows what the engagement is for and what they're expected to bring.
Live: LUMEN methodology (90 min) + applied work (~2h async)
The LUMEN brief format introduced: Learner, Use case, Metrics, Evidence, Next gate. Each participant identifies their candidate use case and starts drafting the L and U sections against their real workflow. Applied work between sessions is to finish the first draft of the brief.
Mentor 1:1 (30 min per participant) + async build
Each participant reviews their LUMEN draft with the embedded mentor. The mentor pushes back on anything vague (success metrics that are not measurable, use cases that are demo theatre, evidence claims that are not auditable). By end of day, the brief is sharper.
Live: LUMEN brief reviews (90 min, cohort + founder) + applied work
Two or three participants present their full LUMEN brief. Founder gives direct feedback in front of the cohort. Everyone sees what a strong brief looks like, what a weak one looks like, and why. Applied work is to finalise the brief.
Brief locked + sponsor-visible artifact log (async)
Each participant locks their LUMEN brief and posts it to the shared artifact log. Sponsor has visibility but is not required to review. Week one closes; participants enter the weekend with the brief done.
What stays the same across cohorts. The cadence, the LUMEN methodology, the mentor 1:1 mechanic, the artifact log, the week-4 sponsor report.
What customises per cohort. Which exercises run against which stack (Microsoft Copilot M365 vs. Anthropic vs. Google vs. OpenAI), which mentor is matched, the verification primitives emphasised (e.g. conformal prediction depth for FS cohorts, brand-voice gate for PS / hospitality cohorts).
Time commitment, per participant
Total weekly commitment is calibrated for senior knowledge workers with a real day job. Senior associates and above can run it without disrupting their pipeline. Junior team members tend to spend the upper end of the range.
One main session (90 min) plus the mentor 1:1 (30 min). Recordings available for makeup if a participant misses live.
Between sessions. Done on your real data and infrastructure. This is the part that compounds; it is not optional.
Across all twelve weeks. Higher at the start (foundations) and during capstone preparation (weeks 10–11).
The mentor and sponsor cadence
One mentor, embedded for the full 12 weeks.
Senior practitioner running production AI at a real company, matched to your sector and stack. Weekly 1:1 with each participant (30 min). Available between sessions in the cohort Slack / Teams channel for blockers. The mentor does not handle commercial conversations; that is the founder. The mentor's job is the work.
Three written reports and one mid-engagement pulse call.
Week-4 report (after Phase 1 close): what was scoped, what's locked, what's been learned. Week-9 pulse (30-min call with HoP, not the founder): is the engagement on track, any yellow signals, recalibrations. Week-12 final report: what shipped, what didn't, the capstone outcome, the operating-model handover. All reports are written by the founder, 2 pages, plain English, no deck.
Deliverables & feedback
What gets produced and reviewed across the engagement.
| Artifact | Who produces it | Who reviews | When |
|---|---|---|---|
| LUMEN brief | Each participant | Mentor + founder | Week 1–2 |
| Verification design | Each participant | Mentor | Week 2–3 |
| Governance log | Each participant | Mentor + (optional) sponsor | Week 3–4 |
| Week-4 sponsor report | Founder | Sponsor | End of week 4 |
| Applied build artifacts | Each participant, on real data | Mentor (1:1) + cohort (live) | Weeks 5–9 |
| Operating brief | Each participant | Mentor + founder | Weeks 10–11 |
| Capstone artifact | Each participant | Sponsor leadership at demo | Week 12 |
| Week-12 sponsor report | Founder | Sponsor + (optional) management committee | End of week 12 |
The capstone
Week 12. Each participant demos the system they shipped. The audience is the sponsor plus any leadership the sponsor invites (CIO, CRO, partners, management committee). The demo format is: 5-minute context, 15-minute walkthrough including the verification gates, 10-minute Q&A. The artifact is real, production-grade or near-production, on your stack, not a slide deck.
If the capstone does not ship at week 12, we have done the engagement wrong. The Bootcamp is structured so the ship date is hit. The few times it has slipped in our engagements, the slip was less than two weeks and was sponsor-driven (priority change at the buyer's end), not Cohorte-driven.
After the engagement closes
Three things travel out of the engagement:
- The shipped system. Owned by you. Built on your stack. Cohorte does not retain code, model weights, or data.
- The operating brief and governance log. Licensed to your firm for internal use indefinitely. The artifact that scales the practice from the cohort to the wider team.
- Engine Room community access for every graduate, for 12 months. The peer network of operators running production AI at other firms.
Other engagement shapes. The Pilot (4 weeks) is a compressed version of week 1 only, scope locked, verification designed, recommendation memo. The AI Readiness Program (3–6 months) inverts the shape, assessment-first, then training, then playbook, then handover. The Curriculum License is async with quarterly live Q&A. Each gets its own page; this one is the Team Bootcamp template because it is the most common.