Section 02 of 12 / Answers and routes
Do-it-yourself game plans
Decide whether your team can run a controlled pilot itself, and how to stage it.
- Pages
- 18
- Sources cited
- 8
- Evidence date
- Oct 20, 2026
Pilot schedule
A controlled pilot, stage by stage
Stage
Weeks 1–2Weeks 3–4Weeks 5–6Weeks 7–8Then
Stage 1 · Exit evidence
Document manifest, timing sheet, agreed task definition and a list of known failure cases.
- Stage period
- Decision stage
- Open-ended
- Ticked complete
0 of 8 stages ticked
Editorial planning estimate only, assuming usable documents, available reviewers and no complex procurement or integration. These stages suggest a 4–8 week narrow pilot, not a typical market duration.
Questions in this section
- 01Can we build and evaluate a useful pilot ourselves?Yes, if you have an accountable estimator, permission to use the documents and someone able to maintain the chosen workflow.Read →
- 02What should the first scope include?Use equipment-schedule extraction as the first scope, with a one-page charter that sets the problem, inputs, output, exclusions and stop condition.Read →
- 03Who needs to participate?Who to involve before the pilot starts, the question each person resolves and the document to obtain from them.Read →
- 04Stage 1 — How do we establish the baseline?Gather complete completed bids, including the difficult versions that the team normally avoids showing in a demonstration.Read →
- 05Stage 2 — How do we define the data so the extraction is useful?Use a controlled schema that preserves original values and units before any normalised fields are added.Read →
- 06Stage 3 — Should we configure software or build a component?Ask the incumbent vendor to demonstrate the narrow requirement before commissioning custom development.Read →
- 07Stage 4 — What data controls should exist before the first upload?Use a project-specific workspace with minimum necessary access.Read →
- 08Stage 5 — How do we build a useful test set?Separate examples used for development from the holdout set.Read →
- 09Stage 6 — How do we measure performance fairly?Run manual and assisted work on comparable packages.Read →
- 10Stage 7 — When can we connect the pilot to live systems?Only after offline output is useful, access has been reviewed and failed transactions can be recovered.Read →
- 11Stage 8 — How do we move to controlled use?Run in shadow mode: the normal estimate remains authoritative while the team evaluates assisted outputs.Read →
- 12What timetable should we plan around?Editorial planning estimate only, assuming usable documents, available reviewers and no complex procurement or integration.Read →
- 13What should the cost worksheet include?Planning assumptions allow $5,000–$20,000 for a narrow pilot and $25,000–$100,000+ for production integration.Read →
- 14What does the final go/no-go pack contain?A seven-point checklist for the final decision, from the before-and-after workflow to a documented exit route.Read →
Reference
Research quality
Self-assessed scores out of 10, not supplier ratings or external validation.
Buyer fit9
Depth9
Search terms9
Sources8
A usable end-to-end pilot method with documents, owners, costs and stop gates. It is not a delivered software implementation or a tested contractor timetable; technical integration design must be adapted to the actual systems.
Read the assessment →