How do AI engines answer questions about your business?

What is tracked

For each business Anserra keeps a list of prompts — the questions a customer would ask an assistant, like “best pediatric dentist near Cedar Park” or “emergency plumber Austin open now”. Each prompt is sent to every base engine several times per run, because AI answers vary between runs.

• Mention: is the business named in the answer at all
• Position: first, second, third, or later among the businesses named
• Accuracy: are the address, hours, phone and services stated correctly
• Competitors: who is named instead of you, and how often
• Sources: which pages the engine cites for its answer
• Call to action: does the answer give a link, phone, or a way to book

Engines

Base engines on every plan: ChatGPT, Perplexity and Google AI Overviews. These three cover the bulk of AI-assisted search in the US and UK. Claude and Gemini are available as add-ons, priced per business, for customers who want them.

Answers are collected through the providers' official APIs, not by scraping chat interfaces. That keeps the data stable and within each provider's terms.

How results are shown

Every number comes with a confidence interval. If your mention rate reads 40% ± 12%, that is the honest range given the number of samples taken. A change is reported only when it moves outside the previous interval, so you are not chasing noise.

History is kept per prompt and per engine, so you can see whether a fix moved a specific question, not just an average.

Cadence and limits

Paid plans run weekly. Each plan has a hard monthly cap on engine responses, shown in your dashboard. Caps are there so that pricing stays predictable for you and for us; you are never billed for overages without asking.