Open-source methodology

AI Enablement Sprint — deploy AI in 90 days.

6 phases. Built on 51 documented enterprise deployments. Same method behind every one of our offers.

TL;DR

→ 6 phases: Discover · Prioritize · Pilot · Production · Scale · Govern
→ Sourced on Stanford 51 deployments, MIT NANDA, Menlo, Deloitte
→ Applied in all 4 packaged offers (CHF 8K → 120K)

See offers

6 principles behind the method

Painkiller, not vitamin

Users who are drowning adopt naturally. Others resist.

Sponsor who persists past the first failure

61% of successes include an earlier failure. Same sponsor.

Start small, controlled, measured

73% of successes start as "an experiment". 80% accuracy = go.

Fix the process before applying AI

AI amplifies whatever process it is applied to (RAND).

Feedback loops beat launch dates

Ship imperfect and iterate weekly. Launch dates are vanity.

Buy infra, build business logic

Purchased AI solutions succeed 2× more than internal builds (MIT NANDA).

The 6 phases

Discover → Prioritize → Pilot → Production → Scale → Govern

Discover

1 week

Map reality, not perception. Shadow AI, data readiness, stack audit.

→ Real usage map (survey + 5-8 interviews)→ Data readiness by function (green/orange/red)→ 20-40 candidate use cases

Anti-pattern — Discovery that drags on 3 months. 1 week gets you 80%.

Prioritize

3 days

From 20-40 candidates to 3 targets. Sponsor signs off.

→ 90-day roadmap (3 use cases, costed)→ SMART success criteria per case→ Governance charter v1

Anti-pattern — Keeping 5-7 use cases "to not offend". The sponsor must cut.

Pilot

2-4 weeks

One use case in production (5-10 users), measured before/after.

→ Use case live for pilot users→ Baseline vs outcome measured→ Go/no-go scale recommendation

Anti-pattern — Scope creep in W2. Additional requests go to v2 backlog.

Production

2-4 weeks

Scale 10 users to 100% without breaking adoption.

→ 100% of users trained (role-specific)→ Prompt library documented→ Adoption dashboard live

Anti-pattern — Go-live without training = 30% adoption ceiling (Stanford).

Scale

Continuous

Use cases 2 and 3 live. Internal AI Lead takes over.

→ 3 use cases in production→ Autonomous AI Council→ Next 90-day plan owned internally

Govern

Transverse

Keep the system alive. 2-page charter, not 200.

→ AI charter signed company-wide→ Monthly AI committee (1h)→ Use case registry + watch

Sources

Stanford — 51 Successful Deployments ↗MIT NANDA — GenAI Divide 2025 ↗Menlo — 2025 State of Gen AI ↗McKinsey State of AI 2025 ↗RAND — Root Causes of AI Failure ↗

Run this method in your company

30-min qualifying call. Honest answer on fit.

Or read the 5+1 Layer Framework (the destination, this page is the path).