Engagement format
An AI pilot on one real workflow
A bounded AI pilot with a baseline, control set, failure profile, and a go / reshape / stop decision before committing to a platform.
Describe the taskProblem framing
What must be resolved
A broad AI initiative lacks one task, a baseline, and an error profile, so a demonstration cannot support a decision about further investment.
Fit
Suitable for
For a team ready to test a hypothesis in real work before committing to a platform programme.
Deliverables
What remains after the engagement
- Prototype
- Baseline
- Benchmark cases
- Evaluation report
- Go/no-go decision
Process
How the work proceeds
- 01Scope
- 02Baseline
- 03Build
- 04Task test
- 05Decision
Scope
From input to a decision
When it is needed
A pilot is useful when a hypothesis must be tested on one real workflow before committing to a platform. It is not a demo for every possible task or a promise of scaling without evidence about use and failures.
Input material
Inputs are one task, one user group, a baseline, a control set of positive and negative cases, and a tolerated failure. The team also names the decision owner and the condition that will trigger a stop or reshape.
What we do
A bounded prototype and comparison with the current workflow are built. The result, failure profile, uncertainty, and user burden are checked. Test data leads to a decision rather than a selection of only successful examples.
Artifact and outcome
The result is a prototype, baseline comparison, failure register, evaluation report, and a go / reshape / stop decision. The material lets a team see what the test supports, what it does not support, and what must be handed over.
Acceptance criterion
Acceptance requires the control set, comparison method, tolerated failures, test outcome, and decision about the next step to be documented and reviewable by the workflow owner.
Boundary of responsibility
A pilot is not an ROI guarantee, an implementation timetable, or a substitute for change management. If the task, baseline, or safety condition is unavailable, reshape or stop can be the honest decision.
Decision record
A pilot decision record joins the baseline, control set, failure profile, and maintenance owner to one choice: continue, reshape the hypothesis, or stop. It is not an automatic ‘go’; an honest decision may show that the material or test conditions are still insufficient.
Boundaries
Risks checked early
- Scope too broad
- No baseline
- Demo instead of a real task
Evidence path
Existing material that demonstrates the scope of work
FAQ
Practical questions
What makes a pilot successful?
A predefined observable improvement with an acceptable failure profile.
How is a pilot different from a successful demo?
A demo shows a possible result; a pilot compares one real workflow with a baseline and control set, including critical failures. Its outcome is a recorded go / reshape / stop decision, not a selection of the best example.
What if the pilot does not meet its criterion?
A failed result is still useful when the failure profile and control set are retained. The team can reshape scope, improve sources, or stop; it should not hide the outcome by changing the criterion after the test.
Scope and cost
Evidence before an estimate
I do not publish a fictional ‘from’. Scope and estimate follow a review of the workflow, material access, risk, expected artifact, and acceptance criterion.
Start by describing the workflow
The first response will identify whether this format fits and what is needed for an honest scope.