A countertop installer should measure AI search impact by combining two habits: asking every new lead how they found the business (and logging answers that mention ChatGPT, Gemini, or "an AI search"), and periodically asking those same assistants directly what they say about the business by name and by category. Standard website analytics miss a large share of AI-driven visits, so the intake form and manual assistant checks become the real measurement tools.
Why AI referrals are hard to see in standard reports
Web analytics tools were built to track clicks from search engine results pages and social platforms, not conversational answers. When a homeowner asks ChatGPT or Gemini "who installs quartz countertops near me" and the assistant names a business without a clickable link, that interaction never shows up as a session, a referral source, or a keyword in Google Analytics. The customer simply calls or walks in, and the visit gets logged as "direct" or "unknown," even though an AI assistant is what actually generated the lead.
Perplexity and some AI Overview results do include citation links, so a fraction of that traffic can show up as a referral from an unfamiliar domain. But a countertop installer relying only on analytics dashboards will systematically undercount AI-driven inquiries, sometimes missing them entirely. That gap is exactly why the intake conversation with new customers matters more for this channel than it does for traditional search traffic.
Questions to add to your intake form
An intake form or phone script that only offers "Google," "Facebook," "referral," and "other" will bury AI-driven leads inside a vague "other" bucket. Countertop installers need a specific option and a follow-up question that captures the assistant name and, ideally, what the customer asked it. This turns a blind spot into a trackable data point within weeks.
Practical additions include a checkbox for "AI search / chatbot (ChatGPT, Gemini, Perplexity, etc.)" alongside the usual referral options, and a short open-text field asking "What did you search or ask to find us?" Front desk staff and estimators taking phone calls should also be trained to ask "how did you hear about us" in a way that invites a real answer rather than a reflexive "Google," since many customers describe using an AI assistant only when prompted with a follow-up like "was that a regular search or did you ask something like ChatGPT?"
How to run periodic assistant checks
Beyond tracking what customers report, a countertop installer can directly test what AI assistants say about the business by asking the same questions a prospective customer would. This means opening ChatGPT, Gemini, Perplexity, and Google's AI Overview results and typing prompts such as "best countertop installer in your city," "who installs granite countertops near your neighborhood," and "compare countertop companies in your city." The goal is to see whether the business gets named, what gets said about it, and which competitors show up instead.
These checks should happen on a recurring schedule, not once. Assistants pull from review sites, business directories, the company website, and other public sources, and what they surface can shift as that underlying information changes. Running the same set of prompts every few weeks, from different devices or accounts when possible, gives a countertop installer a consistent read on whether their visibility is improving, stagnant, or slipping behind competitors who are actively managing their online presence.