Scorecard · No. 10 / 17
How should you evaluate AI visibility software without trusting vendor scores?
Last updated
Compare definitions and evidence before comparing numbers. Public product documentation can identify questions to ask, but it does not establish that a feature works correctly in your account, is included in your quote or is suitable for your category.
| Evaluation example | What the supplier’s documentation says | What the buyer should verify |
|---|---|---|
| Profound | Describes browser-based answer capture and raw-data CSV export | Which surfaces and settings are covered; what the export contains; how omissions, retries and classification changes are handled |
| Scrunch | Describes a combination of browser automation and official APIs, with comparison against directly collected responses | Access route per surface; evidence for the comparison method; whether route changes break historical comparability |
| Ahrefs Brand Radar | Its September 2026 metrics documentation calculates “impressions” using Google-based search volumes and defines share of voice relative to tracked brands | Whether stakeholders understand that this is not a direct count of people seeing an AI answer; exact competitor and weighting rules |
| Semrush AI Visibility Toolkit | Describes a score combining topic coverage and mention consistency; different reports have different data sources and refresh schedules | Which report is being sold, its denominator and update schedule, and whether custom prompt tracking is being confused with database-wide discovery |
Supplier statements above come from their own documentation, not independent validation. (tryprofound.com)
A Profound vs Scrunch AI comparison should therefore examine collection fidelity and evidence access. A Semrush vs Ahrefs AI visibility comparison should examine metric construction and whether the product’s discovery dataset answers your specific buying question. Neither comparison can be resolved by placing proprietary scores side by side.
For scope and pricing, inspect engine add-ons explicitly. For example, OtterlyAI’s August 2026 help page lists some engines as included and others as add-ons. Treat this as a reason to request a complete configuration quote, not as an endorsement or a lasting package description. (help.otterly.ai)
Do not award extra points merely for more engines or prompts. Award them when the additional coverage resolves a documented buyer uncertainty.
Sources cited on this page
- tryprofound.com www.tryprofound.com/features/answer-engine-insights?goal=grow-revenue
- help.otterly.ai help.otterly.ai/pricing-of-otterlyai?utm_source=openai