AI Visibility (GEO)
How AIEO.ee measures your brand across AI answer engines.
GEO (Generative Engine Optimization) is the practice of getting your brand mentioned and cited by AI answer engines. AIEO.ee measures this daily.
Core metrics
| Metric | What it answers |
|---|---|
| Mention rate | In what share of AI answers does your brand appear at all? |
| Citation rate | In what share of answers is your domain cited as a source? |
| Share of voice (SOV) | Of all brand mentions across your prompts, what share is yours vs competitors? |
| Citation breakdown | Which domains do engines cite — and why is it not you? |
Engines
AIEO.ee tracks 10 engines. Three of them support a live AI-search backend (ChatGPT via OpenAI web search, Perplexity via the direct Sonar API, Google AI Overview via Gemini grounding). Whether a given run is live is decided per run:
- Live — the backend key is configured AND the call returned real search
evidence. Citations are genuine. Each run is labeled
data_source: live. - Sampled — every other case: engines without a live backend, a live
engine whose key is not configured, or a live call that returned no search
evidence (automatically demoted). An LLM reproduces what it believes the
answer engine would say. Labeled
data_source: sampledeverywhere in the UI; treat those numbers as a directional signal, not a measurement of the consumer product.
The engine matrix, colors and live-capability flags are defined once in code
(packages/core/src/engines.ts) and rendered from that single source on the
marketing site and in the product; the per-run label lives on each run record.
Prompts & scheduling
You track prompts — the questions your customers ask AI. Each prompt has a
category, an engine list and a cadence. The worker runs prompts daily; every
prompt × engine execution costs 1 credit, is idempotent per day
(re-runs replace the same-day result) and refunds automatically on failure.
Competitor benchmarks
SOV is computed against the competitor brands you configure for the project, so "who is winning AI answers in my category" is one chart, not a research project.