How to Monitor What AI Chatbots Say About Your Company

Short answer: write down the 8–10 questions your buyers actually type into AI, ask them every week in the same way, note whether your company is named, who is named instead, and which websites the AI read before answering. That is how you monitor what AI chatbots say about your company — and it takes about 20 minutes a week by hand.

The rest of this guide shows you the exact method, what we see when we do it for our own businesses, and the mistakes that make people think “AI hates us” when it simply never heard of them.

Why one check is never enough

AI answers move. We track one of our own sites every few days. On 28 September Google's AI named it in 3 of 8 buyer questions. On 2 October, same questions, it named it in 0 of 8. Nothing changed on the site. The AI just picked different sources that day.

If you check once and see nothing, you panic. If you check once and see your name, you relax. Both are wrong. The trend over weeks is the signal; one answer is noise.

What we see tracking our own companies every week

We run several online businesses — automotive data, e-commerce, publishing — and we check them all with the same method. Three things show up again and again:

• AI rarely reads your homepage. It reads directories, review sites, comparison articles and forums. If you are not on those pages, you are invisible, however good your site is.
• Brand questions and buyer questions are different worlds. Ask “Is Acme reliable?” and AI will talk about Acme. Ask “best tool for X?” and Acme is gone. Only the second kind brings new customers.
• AI can confuse you with something else. One of our brand names is also a networking command. For weeks the AI explained the command instead of the business. A clear About page and company facts fixed that.

The weekly method, step by step

1. Write your question list. 8 buyer questions that do not contain your name (“best X for Y”, “how do I choose X”, “is X worth it”) plus 2 brand questions (“is Acme good?”). Keep the list fixed — changing questions ruins the trend.
2. Ask Google's AI with live search on. It shows the pages it used. Those pages are your to-do list.
3. Record four things per question: named or not, position in the list, who was named first, and the sources.
4. Score only the buyer questions. Brand questions are nice to read but they do not bring anyone new.
5. Repeat weekly, same day, same wording. Put it in a calendar.

A simple scorecard

Turning the notes into fixes

Start with the sources. If the AI cites the same three directories every week and you are on none, get listed there first. It is the cheapest win we know.

Then answer the lost questions on your own site. One clear page per buyer question, written in plain English, with the direct answer in the first lines. See how we structure ours on the how it works page.

Then fix the entity. About page with real facts, the same company name everywhere, schema markup. Boring, but it is what stops the “networking command” problem.

What a tool adds (and what it doesn't)

Everything above works by hand. A tracker like ListNetworks does the same steps automatically every week, keeps the history, and turns the sources into a task list. It does not make AI like you — it only tells you the truth faster. If you want to see what that looks like, the sample report shows a full example, and a free report for your own site takes a few minutes from here.