What can you do next without committing to a retainer?

Start with a small, explicitly diagnostic exercise if your commercial question is clear but your baseline is weak. Here is an illustrative specification, not a statistically sufficient sample-size rule:

1. Select six commercially relevant questions and two justified surfaces.
2. Capture each question on each surface on three separate collection dates: 36 planned observations.
3. Preserve exact wording, settings, answers, citations, timestamps and collection outcomes.
4. Count mentions, owned-site citations and explicit recommendations separately for each surface.
5. Record factual errors, unavailable AI features and failed collections separately.
6. Use the results to decide whether a larger, decision-specific measurement design is warranted.

Do not present these 36 observations as 36 independent buyers, total US AI search share, or proof that an intervention worked. For a larger spending decision, require a sampling and uncertainty plan proportionate to that decision; the IAB’s distinction between directional and decision-grade evidence is a useful procurement reference. (iab.com)

What should your completed result say?

Copy this into your internal brief:

Our readiness score: ___ / 45
Our commercial decision: ___
Our lowest-scoring relevant capability: ___
Purchasing gates: passed / not passed; unresolved items: ___
Next action and deliverable: ___
Accountable owner and review date: ___
Maximum authorised external cost and internal effort: ___
Evidence that would make us stop, defer or change direction: ___

The useful result is not a flattering score. It is a smaller, clearer and more defensible next decision.