Product
A measurement platform for AI answers.
Built around how your customers search, reported with the honesty the data allows. Everything on this page is in development; each capability shows its status.
Themes and ranks
Where you rank, theme by theme.
You track themes, each a core need plus the words that change it: place, quality, price or service. Every theme is asked as 15 to 25 company-seeking questions. For each engine you see your average position with its range beside how often you are named, so a single good placement never passes for a strong position.
- Theme view by engine, with weekly movement since tracking began
- Question drill-down: rank, runs and how many answers named no company at all
- A master report per market and language that never blends languages
Lumera Dental
Customers see ranks, competitors and cited sources. Answer text stays internal.
Loading the sample workspace…
Engines and markets
Each engine, market and language on its own terms.
Most engines are measured from the answers a logged-out consumer sees, captured by specialist data providers. Claude is measured through Anthropic's official API with web search, and labelled that way. API answers can differ substantially from consumer answers, so the two collection methods are never averaged together.
| Engine | Collection method | Status |
|---|---|---|
| ChatGPT | Consumer interface | Launch scope |
| Google AI Overviews | Consumer interface | Launch scope |
| Google AI Mode | Consumer interface | Launch scope |
| Gemini | Consumer interface | Launch scope |
| Perplexity | Consumer interface | Launch scope |
| Microsoft Copilot | Consumer interface | Launch scope |
| Claude | Official API | Launch scope |
| Grok | Optional | Launch scope |
Engine coverage is being validated per market and language before launch. An engine that fails validation in a market will not be offered there.
Markets and languages
- UAE · ArabicLaunch scope
- UAE · EnglishLaunch scope
- Saudi Arabia · ArabicConditional
- Saudi Arabia · EnglishConditional
- UK, Australia, New Zealand, US · EnglishUnder decision
Markets and languages are configuration, not code. Gulf workspaces track the Arabic pair first and the English pair alongside it.
Competitor landscape
Everyone the answers put forward.
For every theme, see each company named as an option, whether you track it or not, with its mention rate and average position. Name variants and transliterations merge into one row. Regulators and comparison sites that answers only cite are listed with sources, not as competitors.
- Head-to-head comparisons that say “statistically tied” when they are
- New competitors flagged as they appear, added in one action
- Arabic and Latin-script brand names matched to the same company
| Tier | Company | Mention rate | Average position |
|---|---|---|---|
| 1 | Orvia Dental | 69% | 1.5 (1.4–1.7) |
| 2 | Lumera DentalYou | 53% | 1.9 (1.8–2.1) |
| Dentaro | 48% | 2.1 (1.9–2.3) | |
| Maraya Dental Centre | 44% | 2.2 (2.0–2.4) | |
| 3 | Sablestone Clinics | 34% | 2.7 (2.4–3.0) |
| 4 | Kestrel Smile | 21% | 3.3 (3.0–3.7) |
| Nuvel Dental | 16% | 3.5 (3.1–3.8) |
Companies in the same tier are not statistically different from their neighbour.
Cited sources
The pages engines rely on.
See which domains and pages each engine cites: yours, your competitors' and third parties'. Source-gap analysis surfaces the directories, reviews and publications that cite competitors but not you, the places where authentic presence matters.
Explore cited sources in the sample workspaceEligibility audit
Make sure engines can read you at all.
We check only the levers the evidence calls prerequisites. Every finding links to the engine's own documentation and states its evidence level.
- Crawler access
- robots.txt verdicts for each answer engine's crawler, with training crawlers shown separately as your choice.
- Fetch tests
- Status codes, redirects and bot challenges seen by each crawler's user agent.
- Indexing and snippets
- noindex, nosnippet, max-snippet, canonical tags and Bing's NOCACHE and NOARCHIVE.
- Server rendering
- How much of your main content exists in the raw HTML. Most AI crawlers do not run JavaScript.
- Sitemaps
- Valid sitemaps and a sample of URLs checked for errors.
- Agent readiness
- Whether a computer-use agent can find a price, an address and a booking route on your site. With your written consent only.
llms.txt and schema markup are reported as housekeeping, never as visibility levers: the evidence does not support that.
Crawler access
- PassOAI-SearchBot, ChatGPT-UserAllowed by robots.txtEngine documentation
- FixPerplexityBotBlocked by a wildcard Disallow ruleProven prerequisite
- PassClaude-SearchBot, Googlebot, BingbotAllowed by robots.txtEngine documentation
- InfoGPTBot, ClaudeBot, Google-ExtendedTraining crawlers blocked: your choice, not a visibility issueShown separately
Fetching and rendering
- FixBot challengeA challenge page is served to 2 of 8 answer-engine user agentsProven prerequisite
- FixServer-rendered contentTreatments pages: 31% of main text in raw HTML (threshold 50%)Proven prerequisite
- PassIndexing and snippetsNo noindex or nosnippet on key pagesEngine documentation
- PassSitemapsValid; 50 sampled URLs return 200Engine documentation
Agent readiness
- FixFind a treatment pricePrice not publishedAgent test
- FixBook a consultationBooking form requires JavaScript-only widgetAgent test
- PassFind location and hoursCompleted in 4 stepsAgent test
Housekeeping
- Infollms.txtNot present. No evidence it affects visibility; does not change the scoreInformational
- InfoSchema markupLocalBusiness present. Reported for rich results, not as a visibility leverInformational
Around the measurement
Outcomes
Connect visibility to traffic, honestly.
Import GA4, Search Console and Bing's AI Performance reports to see AI-referred sessions and conversions by landing page, next to the pages engines cite. Every outcome figure is labelled a floor: some AI traffic arrives without a referrer, and Google's AI Overview clicks cannot be separated in GA4.
Getting started
Start with a look. Go deeper when it matters.
Sign-up starts with your domain and a work email. An entry plan is planned so any company can see where it stands: three themes and 15 questions collected each Saturday, shown as counts. Paid plans add full ranges and ranks, more themes, engines and markets, deep research, the eligibility audit and first-party data.
Plans and prices are being finalised and will be published before launch.
Personal AI agents
Agents that act, measured with care.
Assistants such as OpenAI dots, Instinct, Hark and Grok Bot choose a business and book or enquire. None offers a public API today, so no one can measure them directly. We plan three honest steps: simulated tasks through the closest public agent API, always labelled “simulated with” the stand-in; a test of whether an agent can complete an enquiry on your own site; and a watchlist ready for when access opens.
SimulatedAPI and MCP
Your data, wherever you work.
The web app is built on the same public API you can use: read and write endpoints, signed webhooks for alerts and reports, and an MCP server with read tools so assistants can query your results, with labels and provenance attached to every figure.
What we won't sell you
Some things don't belong in a measurement product.
- Guaranteed rankings or citations
- Single-run “rank #1” screenshots presented as fact
- One blended score across languages, markets or engines
- Hidden text, prompt injection or AI-only page versions
- Mass-generated pages, listicles or review campaigns
- llms.txt or schema sold as a visibility lever
- “Prompt volume” data from consumer panels
See it on your own themes.
Early access starts with a pilot in the UAE, in Arabic and English.
Request early access