FAQ

Short answers to the questions we get most. Each one links into a deeper article when there's more to say.

Why do my results differ from what I see in my own ChatGPT?

Personalisation. Your browser session has months of conversation history, location bias, and account-level preferences shaping the answer. We query each engine through a controlled production pipeline with none of that history attached, so what you see in the dashboard is closer to "what a typical, fresh user would see right now". See AI engines.

Why isn't my brand mentioned even though my domain is cited?

AI citations and source URLs are different signals. AI can list a URL on your domain (counts toward source URLs) without naming the brand in the prose (no AI citation of the brand). Read AI citation vs. source URL for the difference.

How fast does the score update?

The Overview hero refreshes after every pulse run. The cron wakes daily at 08:00 UTC, but each monitored prompt is only due once a week — the exception is a Hero Prompt, which is pinned to daily. Manual scans update it immediately. Trend charts add a new point per scan.

Why does GEO Score on Pro vs Business look different for the same data?

It doesn't — the formula is identical across paid plans. Different numbers between tiers usually mean a different scope: Business plans scan more queries and may include more engines per scan if you've allocated them, which changes what's included in the average.

Can I track multiple domains?

Multi-domain UI is on the near roadmap — today every plan has a hard cap of 1 active tracked domain per account, enforced at the database level. Business and Scale plans will unlock 3 and 10 domains respectively when the domain-switcher UI ships.

How many AI engines do you measure?

Up to five, and which ones run depends on your plan: Free measures Perplexity; Starter adds ChatGPT; Pro adds Google AI Mode and Gemini; Business and Scale add Google AI Overview. These cover the great majority of B2B buyer-discovery traffic in 2026, and we re-evaluate the line-up every quarter. See AI engines for the full ladder.

What counts as a "manual scan"?

A scan triggered from the product UI (Run scan, on-demand audit) as opposed to one started by a cron. Manual scans count toward the plan's manual-scan allowance — see Pricing.

What's the difference between Google AI Mode and Google AI Overview?

They're two different surfaces, and we measure them as two different engines. AI Mode is the dedicated chat tab on google.com (always answers) — tracked on Pro and up. AI Overview is the AI panel above the classic search results (Google chooses when to render it) — tracked on Business and Scale. Full explanation: Google AI Mode and AI engines.

Can I export my data?

Yes — Pro, Business, and Scale plans expose CSV and JSON exports at /api/export/scans, /api/export/citations, and /api/export/geo-history. Pass ?format=csv or ?format=json. The window is capped at 365 days and 10 000 rows per file; for larger pulls, run multiple windowed requests. Free and Starter plans see a Settings → Billing upsell.

How does competitor auto-discovery work?

Every scan reads the engine-native citation arrays plus brand mentions in the prose, runs a blocklist + subdomain prefix + TLD whitelist, and upserts the surviving hostnames into your competitors table. Read Competitor tracking for the filters and tier limits.

Why don't you write the copy for me?

Because AI-generated marketing copy is currently slop, and pushing slop to your site would damage the trust signals we're trying to improve. JSON-LD and FAQ schema are structural — those we generate happily. Everything else stays your editorial decision.

Where can I see the actual AI answers?

The Citation excerpt search on the Scans tab. Search any term, expand any hit, the full answer is inline with your search terms in yellow and your own brand in cyan.

How accurate is the mention quality classification?

The deterministic pass handles obvious patterns (X is the best, X | Static sites & Jamstack, listicle position). Long-tail or ambiguous cases route to a small refine call. Quality is reasonably stable across re-runs but isn't deterministic — we publish the buckets in GEO Score so you can audit any individual call.

Is my data shared anywhere?

No. Data is row-level-security scoped to your user in our Supabase project. Calls out to our AI engine and data providers send the query text only — never your competitor list, brand internals, or account information.