Disclosure: I built the Apify Actor in this post. Written with AI assistance.
A buyer types "best CRM for small business" into ChatGPT or Perplexity and gets three or four brand names plus the pages behind them. Is yours one of them, or is it your competitor? Most teams have never checked, because asking the chatbot by hand once tells you almost nothing.
Here is a first audit you can run in about five minutes with AI Brand Monitor on Apify. No code, no API key.
The audit
- Open the AI Brand Monitor page. The form is prefilled with an example (hubspot.com and five CRM questions).
- In Your brand domain, replace the example with your own domain, such as
example.com. - In Queries to track, replace the five prefilled questions with five that your buyers really ask. Good patterns: "best [category] for [audience]", "[your brand] alternatives", "[your brand] vs [competitor]", "is [your brand] worth it".
- In Competitor domain(s), add three or four rivals.
- Leave Engines to query as prefilled (Perplexity and ChatGPT) and Samples per query at 1. Click Start.
Not ready to pay anything yet? Tick Free sample run first: it runs your first 3 queries at 3 samples each on the included engine with a result fee of $0.
What you get back
One dataset row per AI answer. The fields to read first:
-
brandCited: did the engine cite your domain or mention your brand in the answer text? -
competitorsCited: which of your competitors showed up instead. -
citationShare: your share of the citations against your competitors. -
answerText: the full answer, so you can read exactly how you were (or were not) described.
If brandCited is false on most rows and a rival is named on all of them, you have your answer, and a clear gap to close.
What it costs
You pay per AI answer, called a citation record: $0.05 on the free Apify plan, less on paid plans (down to $0.035). The audit above is 5 questions x 2 engines x 1 sample = 10 records = $0.50. On a free plan ChatGPT is measured on the first 3 questions per run and skipped questions are never billed, so that run comes to 8 records = $0.40. If you want a hard ceiling, set Max total charge before you press Start.
Make the result trustworthy
AI answers vary from request to request, so one sample is a snapshot. Raise Samples per query to 3 and the Actor also reports a stability score for each source: how consistently it was cited across the repeats. The Actor's docs recommend ignoring sources below 0.5.
One more detail worth knowing: the Actor asks ChatGPT through OpenAI's API with web search, then records the answer and its sources. It is a measurement you can repeat and compare, not a screenshot of your own chat history.
Next step
Run the five questions that matter most to your sales today. Whatever the audit shows, you will know which brands the AI engines put in front of your buyers, and you will have the source pages they used.
Run your first AI visibility audit, then save the input as a task so you can re-run it next week and see what moved.
Top comments (0)