Methodology
Test patterns, not screenshots.
AI answers change over time, so a single snapshot is not enough. We run structured, repeated diagnostics, classify the sources shaping answers, verify claims, and translate findings into a practical plan.
How it works
Seven steps, repeated over time.
- 1
Define the buyer prompt universe
Prompt sets built around category, comparison, alternative, local, and objection questions.
- 2
Run repeated measurements
Test across selected AI systems over a window to observe patterns, not isolated outputs.
- 3
Extract mentions and citations
Capture whether you appear, which competitors appear, which domains are cited, and what the answer claims.
- 4
Classify source types
Group sources into owned, earned, reviews, directories, social, partner, competitor, or unknown.
- 5
Score visibility and influence
Separate basic presence from answer influence.
- 6
Verify claims
Flag unsupported, outdated, or incorrect AI statements for correction.
- 7
Translate findings into actions
Content, source, schema, and messaging fixes tied to observed gaps.
Measurement
Clear metrics, tied to action.
We report the signals that influence AI answer visibility in terms executives can act on.
AI visibility score
Trending up across the last six re-audits.
AI visibility is measurable and influenceable, not fully controllable — results vary by model, prompt wording, location, source availability, and time.
Deliverables
A stack of executive-ready artifacts.
Built for decisions and execution — not raw model dumps.
Executive Scorecard
Prompt Performance Matrix
AI Answer Evidence Packet
Citation Map
Accuracy / Hallucination Register
Content Gap Plan
Vendor Accountability Notes
30/60/90-Day Action Roadmap
Built on research, not guesswork.
See the external references that inform how we test and measure.