Questions
30 per category: 12 discovery ("best AP automation for NetSuite"), 6 comparison ("Stampli alternatives"), 6 fit ("AP software with 3-way matching") and 6 fact questions (price, who it is for, integrations, security, implementation time, reviews).
Engines
ChatGPT with search, Perplexity, Google AI Overviews and Grok, each queried logged out with a US locale. Every report names the engines actually queried. We never fill in an answer we did not collect.
Variance
AI answers change from run to run, so each question is run 3 times per engine in separate sessions. We report mention rates with their sample size and keep the full answer text and cited URLs. Small samples are labelled directional.
Scoring (0 to 100)
| Component | Points | How it is measured |
|---|---|---|
| Share of voice in discovery questions | 35 | Runs that name you, divided by all discovery runs |
| Position | 15 | Average rank when you are named (1st earns full points) |
| Comparison presence | 15 | Named in "alternatives" and "vs" answers that do not name you in the question |
| Accuracy | 20 | Fact answers graded against your own pages: correct, partial, wrong or absent |
| Source control | 15 | Share of citations from your own domain, plus technical readiness |
Limits
Answers vary by user, location and personalization. The audit measures a logged-out baseline. A mention rate is not revenue, and we make no traffic or revenue claims without your own analytics.