What should happen when a pilot exposes a missing dependency in the task scope?

A model lab scopes a pilot around evidence that professional services engagements received authorised acceptance. The proposed inputs include engagement ENG-2187, acceptance form ACC-0412 and a policy describing which roles can approve the work.

During sample review, an evaluator asks whether the approving role held the required authority on the acceptance date. The form records a role identifier, but the pack contains no assignment history. The pilot issue changes from accepted for testing to blocked by a scope dependency: the intended conclusion cannot be judged from the proposed inputs.

The buyer has a concrete choice. Add generated authority history to support the original task, or narrow the task to checking that an acceptance form exists and contains the required fields. The latter remains testable but measures a different capability.

The consequence of ignoring the dependency would be a misleading evaluation that rewards assumed authority. This scenario teaches a useful scoping discipline: when a required inference depends on missing context, change the input contract or change the claim being tested. Additional document variety would not repair that gap.