Create a vertical workflow map and manual pilot report
Map one industry workflow, run one manual AI-assisted delivery, and prove saved time or metric improvement.
When this is the right output
Best for people with industry access who can reach decision-makers and handle delivery boundaries.
Prepare before you start
- One familiar industry
- Ten practitioner interviews or three real workflow materials
- One measurable metric: time, speed, conversion, or error rate
Build the first version
Map the workflow
Start from material intake and end with final delivery. Mark human judgment points.
Confirm strong pain
Move to pilot only after the same blocker appears at least three times.
Run manually
Do not integrate systems yet. Use AI plus human review once.
Write the report
Report only verifiable metrics such as time saved, errors reduced, or faster response.
Copy-ready templates
Interview questions
Who owns this workflow? How often does it happen? Where is the slowest step? Who pays for mistakes? What result would justify a pilot?
Pilot report structure
Current workflow, input material, AI intervention, review boundary, before-after metrics, next pilot scope.
How to decide the next step
Continue when
- Three practitioners repeat the same pain
- One client provides material for a pilot
- The pilot shows time saved or a metric improvement
Change direction when
- The pain is custom and not repeated
- Material cannot be obtained or anonymized
- No measurable metric exists
Return to the result after you finish this output
In week one, create the workflow map, input checklist, pilot report template, and 3 pain signals.