โ† Back to The GenAI Playbook
MING Labs ยท Framework 03

6-Week POC-to-Profit โ€” Planning Canvas

Use case: ____________________ Owner: ____________________ Start date: ____________________

Use Case

One sentence: what does it do?

Owner

Name + title (one person)

Success Metric

One number that proves value

Kill Criteria

"If ___ by week 4, we stop"
Week 1

Scope Lock

One use case, one owner, one metric. Max 3 people.
Deliverables:
Use case defined
Owner confirmed
Kill criteria set
Data source identified
Can't lock scope? Not ready.
Notes:
Week 2

Data + Architecture

Connect to real data. Build simplest working version.
Deliverables:
Real data connected
Architecture decided
Working skeleton (ugly OK)
โœ“ Functional skeleton on real data
Notes:
Week 3

Build + Test Round 1

Functional prototype. First user testing (3-5 actual users).
Deliverables:
Prototype functional
3-5 users tested
Feedback logged
User feedback:
Week 4

Iterate + Test Round 2

Refined prototype. Second user testing. Set accuracy expectations.
Deliverables:
Iteration complete
2nd test round done
Human fallback designed
No adoption by week 4? Stop.
Notes:
Week 5

Measure

Parallel run with existing process for 5 days. Hard numbers.
Deliverables:
Parallel run complete
Time savings measured
Error rate documented
ROI calculated
โœ“ Hard numbers on value
Results:
Week 6

Decide

One room, one owner, one hour. Scale, Pivot, or Kill.
Deliverables:
Results presented
Decision made
Next steps documented
Decision & rationale:

Baseline (Before)

Time per task today:
Error rate today:
Cost today (annual est.):

Result (After POC)

Time per task with AI:
Error rate with AI:
Projected annual savings:

Economics

POC investment:
Projected annual value:
Payback period:

Week 6 Decision

Scale
Pivot
Kill