Measurement, control 5 of 17
M-05 Measure AI visibility by topics and user tasks
Build representative topic and journey cohorts rather than treating one prompt as the whole market.
- Priority
- High priority.
- Hard gate
- No
- Change class
- READ_ONLY: Inspect authorised evidence.
- Quality bucket
- Measurement and experiment readiness
Why it matters
A single keyword or prompt gives a narrow and unstable picture of how a brand appears in generative answers. People ask about a subject in many ways, from discovery to trust to support, and a measurement set that covers only one of them misses most of that. Topic-level cohorts show where visibility is strong or missing across the whole subject.
Evidence to review
- Versioned prompt library, weighting rules and cohort coverage.
How to implement
- Cover discovery, evaluation, selection, trust, local and post-purchase needs as relevant.
- Keep branded and unbranded prompts distinguishable.
- Set commercial weights before viewing results.
- Do not blend engines or account conditions into an unexplained number.
How UltimateSEO checks this
Automated. A check in the scan sets PASS or FAIL for this control, and reports UNKNOWN or BLOCKED when its input is missing.
States used: PASS, FAIL, UNKNOWN, NOT_APPLICABLE or BLOCKED. UNKNOWN is never shown as a pass. How states work.
Myth guards that reference this control
Sources
Part of The UltimateSEO Framework v4.3, catalogue dated 2026-10-14. Evidence states, control states and the publish gate are explained on the hub.