Illustrative pilot output
A recommendation with its limits visible.
We entered a fictional facade Case into the Web Pilot with emails, a site log, a drawing, a synthetic photograph and a recorded human account. This is an edited public extract of the pilot’s actual report, not a real construction decision or independent validation.
Illustrative report extract
This one-page English extract reflects the analysis of a fictional Web Pilot Case. Select the preview to open the PDF. It is an edited extract, not the complete system export.
Open the English extract (PDF) ↗Illustrative Case · W-06-17 facade closure · 25 September 2026
Do not release the opening for closure yet.
Pilot result: conditionally ready to advise on the hold; not ready to authorise closure.
- Decision at hand
- May the facade at opening W-06-17 be closed before the 16:00 crew slot? What minimum verification is needed if it cannot?
- What the sources support
- The QA hold point remains open. Proposed sketch C is for discussion and has not been formally issued. A site observation records a visible discontinuity at the bracket; continuity on the concealed side has not been verified.
- What remains uncertain
- There is no documented inspection after restoring the barrier and no QA release. The original approved detail was not supplied. Moving the crew to another opening has not been checked for access, materials or programme effects.
- Consequences and options
- Closure may miss its planned slot. Keep W-06-17 on hold while the barrier is checked against the approved detail and the proper QA process is completed. Consider reassigning the crew only after verifying the other opening and its constraints. Cost and wider schedule effects cannot yet be quantified.
- Recommendation
- Ask the person with operational authority to confirm that W-06-17 stays on hold until barrier verification and QA release. If the 16:00 slot cannot be met, agree a crew priority separately. A designer’s opinion does not itself issue a site instruction or approve a client change.
- Human decision
- Rabbit prepared the decision request and its basis. No person’s final decision or subsequent outcome is asserted here.
Evidence formats in this Case
We submitted four .eml email messages, three .txt site and QA records, and two .png images: a discussion drawing and a synthetic photograph. Two attributed human accounts were entered directly into the Case rather than uploaded as files. This describes the input formats, not what the pilot accepted as proof; the photograph was excluded from the recommendation’s basis.
What this example also exposes
The pilot did not treat the synthetic photograph as reliable Case information: its extraction and provenance were insufficient. The human update remained an attributed account, not automatic proof of an inspection. These limitations are part of the result.
This public extract translates and shortens the report; it does not reproduce the underlying emails, image, drawing or full internal audit trail. Read how the method works.