The Audit · Premise to Practice Method
Find the work worth changing before you buy another AI tool.
The Audit uses the opening moves of the Premise to Practice Method. It helps a leader or team name the outcome, map the real work, question inherited constraints, and choose one useful pilot.
The Decision
What should we build first?
Timing
Usually completed in two weeks
Focus
One team or business function
Result
One prioritized pilot and a leadership readout
Most teams can name ten places AI might help.
They get stuck on which idea deserves time, what the work actually looks like, what information it touches, and how anyone will know the change helped.
The audit starts with what the work is for. You leave knowing what to test first, why it deserves the effort, and what has to be true before the work expands.
Outcome before artifact
The decision, result, or change the work exists to produce before anyone chooses a deliverable or tool.
Current work and handoffs
Where the work begins, who touches it, what slows it down, and where quality breaks.
Inherited constraints
Which steps exist because of an old system, missing access, or manual limitation that may no longer apply.
Floor, ceiling, and proof
Where AI can remove steps, where the outcome could be reached differently, and what evidence would prove it.
Enough detail to make the next decision.
The output is built for action. It gives a leader something clear enough to approve, question, fund, or stop.
Work System Map
A plain view of the recurring work, handoffs, decisions, inputs, and pressure points.
Opportunity Scorecard
A ranked set of candidate workflows based on value, risk, readiness, and ability to test.
Pilot Brief
One recommended starting point with an outcome, owner, boundary, quality bar, and first test.
Guardrail Draft
Initial rules for data, review, escalation, stopping, and recovery.
Proof Plan
The examples, measures, and reviewer confidence needed to judge the pilot honestly.
Leadership Readout
A direct recommendation on what to build, what to leave alone, and what decision comes next.
01
Name the outcome
Define the decision, result, or change the work exists to produce.
02
Map the real work
Trace the steps, handoffs, inputs, pain, ownership, and boundaries.
03
Question constraints
Identify steps created by limitations that may no longer deserve a place.
04
Choose the test
Select the strongest starting point and define the proof needed to judge it.
A useful fit
- You lead a function or team with recurring knowledge work
- You have more AI ideas than operating clarity
- You need a credible pilot instead of another general training session
- You can bring safe examples and people who understand the work
A poor fit
- You need production software delivered by an engineering team
- You want a vendor shortlist without examining the work first
- The only available examples contain unapproved sensitive data
- Leadership has already decided the answer and only wants the audit to validate it
What Happens Next
The audit ends with a decision, not an automatic upsell.
You may be ready to build, need leadership support around the pilot, or decide the idea should stop. The point is to know which one is true.