Ranking proof you can put in front of a board.
Every ranking claim CitedLogic makes is backed by an artifact a stakeholder can open, read, and date. The baseline becomes a board-ready market map and execution plan, not a black-box score.
Every finding is a receipt.
A finding without proof is an opinion. Each result in your report carries four artifacts.
Screenshot
The answer as the customer’s device rendered it.
Transcript
The full prompt and response, captured verbatim.
Timestamp
When the session ran, to the minute.
Metadata
Engine, metro, and session context attached to every record.
Every ranking claim has to survive inspection.
The methodology is transparent. The operating system is proprietary.
Premium buyers need enough evidence to trust the service. Competitors do not need the capture mechanics, prompt library, or ranking recipe.
Evidence standard, selected before/after examples, aggregate proof stats, methodology boundaries, and sample report summaries.
Evidence Register, vertical field guides, sample baselines, AI reputation risk briefs, weighted execution maps, and board-ready proof packs.
Raw evidence folders, exact prompt set, market opportunity maps, source-level fix maps, weighted execution plans, and operating mechanics.
Bad captures are quarantined. A screenshot error does not become a public claim. Public proof only ships when the visible receipt supports the movement story.
Metro-local, real-device sessions.
We measure the way a customer in your market actually experiences AI - location-controlled, on real devices, across the engines that matter.
Location-controlled measurement
Each session is controlled to a specific metro, so the answer reflects what a local prospect sees — not a national average and not a server in another region.
Metro-local real-device sessions
Readings come from real-device sessions rather than synthetic API or headless calls, capturing the experience an actual customer’s phone returns in the market.
Across every answer engine
We measure across the engines your customers actually use to decide.
Handled like the record it is.
What we store, how it's protected, and who can touch it — written plainly.
What we store
Audit artifacts (screenshots, transcripts, run metadata) and your account configuration. We don't sell data or share it with third parties.
Encryption
Data is encrypted in transit (TLS) and at rest. Evidence is retained for your reporting window and disposed of on request.
Access controls
Least-privilege access, scoped to the team members who operate your account. Administrative actions are logged.
Security posture
Full security and data-handling details are available under NDA; SOC 2 is on the roadmap and we'll publish status here, not before.
We show our work.
Including where methods disagree — because a number you can't interrogate isn't evidence.
Synthetic-vs-real-device disagreement, disclosed
Where a synthetic reading would have diverged from the real-device result, we disclose the disagreement rate rather than hide it. It’s the clearest proof of why real-device measurement matters.
Benchmarks are labeled as our data
Any benchmark or share-of-answer figure is labeled as CitedLogic data, scoped to the markets and window it was measured in — never presented as an industry-wide absolute.
See the evidence for your own markets.
An AI ranking baseline, delivered with the screenshots, transcripts, timestamps, and weighted execution plan behind every finding. Bring it to the board.