How We Measure Whether AI Search Is Actually Working for You
AI search is new enough that most of the metrics floating around are noise. When someone asks ChatGPT, Gemini, or Google's AI Overviews for "the best plumber near Farmville" or "where to get custom framing in Prince Edward County," the answer they get is now shaping who calls you. The hard part is proving it is working, because these tools do not hand out a tidy ranking report the way traditional search does. So here is exactly what we track — and, just as important, what we refuse to waste your money chasing.
What we actually measure starts with citation presence: are you named or linked when an AI assistant answers a real buying question in your category and area? We run the prompts a Farmville customer would actually type, across the major assistants, and we log whether you appear, whether a competitor appears instead, and how you are described. That description matters — an AI that calls you "a family-owned shop serving the Farmville and Longwood area since [year]" is doing sales for you. One that gets your services wrong is costing you customers.
Second, we watch referral traffic and conversions from AI sources. Assistants increasingly send clicks, and those visitors tend to arrive further along in their decision than a cold search. We tag and separate that traffic so you can see not just that AI mentioned you, but that the mention turned into a form fill, a call, or a booking. A citation that never sends a warm lead is a trophy, not a result.
Third, we track the accuracy and freshness of what the models say about you. In a market this tight-knit, being described wrong travels. If an assistant is quoting outdated hours, the wrong service area, or a service you dropped, we treat that as a defect to fix at the source — your site, your profile, the structured data — not something to shrug at.
- Vanity metric we ignore: a raw "AI visibility score" from a dashboard that never ties to a lead
- Vanity metric we ignore: appearing for prompts no Farmville customer would ever type
- Vanity metric we ignore: mentions in generic national answers that bring you zero local business
- What we care about: named in local buying-intent answers, and warm traffic that converts
The reason this discipline matters here specifically is that Farmville is a real-relationship town — Longwood, Hampden-Sydney, seven counties of people who ask each other and now ask their phones for recommendations. Being the answer the AI gives is the modern version of being the name a neighbor drops. We measure it the same way you would judge word of mouth: not by how loud it is, but by how many of the right people it actually sends through your door. Our AI Search page lays out the full reporting cadence.