How accurate is the free scan?

Every free scan result comes from a real ChatGPT answer recorded at scan time, never a simulation. You can open the receipt behind each result and verify exactly what the engine said.

Accuracy in AI visibility has two parts, and it is worth being precise about both. The first part is whether the data is real. Citably runs your prompts as live queries and stores the full response. When a result says you were cited, the sentence that cited you is in the receipt. When it says a competitor was named instead, their name is in the receipt too. There is no black box to trust.

The second part is repeatability, and here honesty matters: AI engines are probabilistic. The same prompt can produce a slightly different answer an hour later, which means any single scan is a snapshot, not a verdict. This is a property of the engines themselves, not of the measurement.

Two practical consequences follow. Read trends over 60 to 90 days rather than comparing two individual scans, and treat a one-off change as noise until it repeats. Why does my score change between scans covers this in detail.

Also remember the free scan covers one engine. ChatGPT is the highest-traffic answer engine, so it is the right single check, but a brand can look different across Gemini, Perplexity, and Claude. Paid scans across all four give the fuller picture.

Related articles