Standard analytics undercount AI-driven visits because tools like ChatGPT, Gemini, and Perplexity often send traffic with no referral tag, or the customer never clicks a link at all — they just read the AI's answer, get your business name and phone number, and call directly. To measure this accurately, a chimney sweep or repair business needs to combine analytics review with direct customer questions and hands-on testing of the AI tools themselves.
Why standard analytics undercount AI-driven visits
Google Analytics and similar tools were built to track clicks from search results pages and social links, sorting traffic into neat channels like "organic search" or "direct." AI engines break that model. When a customer asks ChatGPT "who can fix a cracked chimney crown near me" and gets your business name in response, they frequently search your name separately, type your website straight into the browser, or just dial the number they were given. None of that shows up as "AI search" in a dashboard — it lands in "direct traffic" or disappears entirely if the customer never visits your site.
How to ask new customers how they found you
Asking every new customer a single follow-up question is the most reliable way to catch what analytics miss. When someone calls to book a sweep, inspection, or repair, a simple "how did you hear about us?" gives you information no tracking software can capture, including whether they got your name from an AI answer rather than a traditional search result or a friend's recommendation.
Train whoever answers the phone or books appointments to log the answer in a simple spreadsheet or your scheduling software's notes field. Listen for specific phrasing: customers who found you through an AI tool often say things like "I asked ChatGPT" or "I looked it up on my phone and it just told me to call you" rather than "I found you on Google." Over a few months, this log becomes a clearer picture of AI-driven demand than any analytics report, because it captures phone calls and walk-ins that never touch your website.
What to watch in referral and direct traffic
Referral traffic and direct traffic in your analytics dashboard hold clues about AI search activity, even though they were not designed to show it. A jump in direct traffic that does not match any offline marketing you have run recently, especially when it lands on your homepage or a specific service page rather than spreading evenly across your site, can indicate that an AI engine has been pointing people to a particular page.
Check your analytics referral report for domains you might not recognize immediately, since some AI tools do pass a small amount of referral data when a user clicks a source link inside an answer. Look for landing page patterns too: if visits are concentrated on your "emergency chimney repair" or "chimney inspection cost" pages rather than your homepage, that suggests people arrived already looking for a specific answer, which is consistent with how AI tools summarize and link to specific pages rather than sending people to a site's front door.
How to test your own business in the engines
Testing your own business directly in ChatGPT, Gemini, and Perplexity shows you exactly what a potential customer sees before they ever call you. Open each tool and ask the kinds of questions a homeowner would type, such as "who does chimney repair near your town" or "best chimney sweep for a cracked flue in your area." Note whether your business appears, what information the engine states about you, and whether that information is accurate.
Run this same test on a phone and a desktop browser, since results can differ. Pay attention to what the engine says about your hours, service area, and specialties. If an AI tool is quoting outdated information or leaving out a service you offer, that is a signal worth acting on, since it means the version of your business being presented to potential customers does not match reality. Repeat this test every so often rather than once, because AI-generated answers can shift as these tools update the sources they rely on.
Setting simple monthly checkpoints
A short monthly review turns scattered clues into a real measurement habit instead of a one-time curiosity check. Set aside time once a month to pull together three things: the "how did you hear about us" log from new customers, a quick look at direct and referral traffic trends in analytics, and a fresh round of testing your business name and services in the major AI search tools.
Write down what changed since the previous month, even in plain language: more customers mentioning AI tools by name, a landing page getting steady direct visits, or an AI engine now naming your business when it previously did not. This monthly checkpoint does not need to be complicated. Its value comes from consistency, because tracking the same three signals every month is what reveals a trend, rather than one lucky month of new customers you cannot explain.
The single next step that outranks everything else this month
Start the "how did you hear about us" log today, before doing anything else on this list. Analytics reports and AI engine tests both matter, but they only tell you what is technically possible or plausible. Direct customer answers tell you what is actually happening in your business right now, and every week without that log is a week of information you cannot get back. Once a handful of new customers mention an AI tool by name, you will know with certainty that this channel deserves closer attention, and every other checkpoint in this article becomes easier to interpret because you will have a real baseline to compare it against.