GEO Monitor: does your brand show up when someone asks an AI?
More and more buyers ask Gemini, Perplexity or an OpenAI model directly instead of searching on Google. The GEO Monitor measures, with real questions and real answers from those engines, whether your brand gets cited — without inventing an abstract metric or promising more certainty than a non-deterministic result allows.
How it actually works
- Set up to 12 questions per language (or use the 5 product-default questions, written for your type of business).
- When you press "Run GEO check", each question runs on Gemini and, when configured server-side, also on Perplexity Sonar and/or an OpenAI model.
- Each answer is checked for whether it cites your brand or domain — the result is "Cited" or "Not cited" for every engine-and-question combination.
- The GEO Score is the percentage of all those combinations that were cited, with a per-engine breakdown, a trend chart over time, and concrete suggestions when the score is low.
The GEO Score, in numbers
- 0-20 — Invisible to LLMs.
- 21-50 — Emerging presence.
- 51-75 — Moderate visibility.
- 76-100 — Sector reference.
It is a trend to track over time, not an official ranking or a fully deterministic number — the same question can vary slightly between runs, because that is how generative AI models work.
What it costs to run
Each run consumes 1 AI credit from your plan — not 1 credit per question or per engine, no matter how many you use in that run. Results stay cached on the server for about 24 hours, so repeating the same check within that window doesn't spend another credit. AI credits are shared with DENVORA SEO's other AI features (Starter 10/month, Professional 50/month, Elite 250/month).
When your brand doesn't show up
It isn't a product error or an AI engine fault — it's a signal that there isn't yet enough consistency between brand, category and geography across the sources the models can see. The suggestions the panel shows include:
- Strengthening brand and domain mentions on pillar pages and in Organization schema.
- Publishing case studies and FAQs with the entity (brand/business) front and center.
- Looking for mentions and backlinks from directories and partners relevant to your market.
- Keeping sources consistent: same name, same domain, same category everywhere.
The three engines, side by side
Gemini and Perplexity Sonar use near-real-time crawlers, so they tend to reflect site changes faster. The OpenAI engine (via Chat Completions) relies on the model's trained knowledge, with no live web search — the same underlying knowledge behind ChatGPT, but without the browsing layer the real ChatGPT app uses, so it usually takes longer to pick up new signals.
Frequently asked questions
Does the GEO Monitor really test ChatGPT?
It tests the OpenAI model via Chat Completions — the same trained knowledge behind ChatGPT, but without live web browsing.
Is the result an official ranking position?
No. It's the percentage of cited answers — an observable trend, not a deterministic position like a SERP rank.
How much does each check cost?
1 AI credit per run, regardless of the number of questions or engines. 24h server-side cache.