How FirstCited measures AI visibility

What FirstCited measures

Search used to return a list. Increasingly it returns an answer with 2 or 3 names in it. FirstCited measures 2 things about that answer:

• Named. Does the engine mention your business by name in its answer to a buyer's question?
• Cited. Does it link to a page you control - your site, your Google Business Profile - as the source for that mention?

We run it across the engines a buyer is most likely to be holding: ChatGPT, Gemini, Perplexity and Google's AI answers. Claude can be added to a run if your buyers use it.

How a baseline run is done

A baseline is a dated snapshot of how the engines answer your buyers' questions, taken before any work starts.

• The queries. We write them the way a buyer types, not the way a marketer would. For a cosmetic dentist in Scottsdale that's things like "best cosmetic dentist Scottsdale," "veneers Scottsdale cost," "who does the best smile makeover in Scottsdale." Enough to cover the head term, the price question, the comparison question and the specific-procedure questions, without padding the set. You see the queries before we run them.
• Logged out. Every query runs in a fresh browser session with no account signed in. A signed-in engine personalizes on your history, and a personalized answer isn't the one your buyer sees.
• New thread per query. Each query goes into its own new conversation. Engines carry context inside a thread, so a second question in the same thread gets a different answer from a first one.
• Same day. Every query on every engine runs in 1 sitting and is logged with the date. Engines change their answers over time, so a baseline only means something with a timestamp on it.
• Logged verbatim. For each answer we record the exact query, the engine, the date, every business named, the order they're named in, and every URL the engine links. Screenshots are kept.

What the result tells you

You get 2 counts: how many of the answers named you, and how many linked to a page you control. We also report who was named instead of you. That list is usually more useful than your own count, because it shows which competitors the engines already trust and which sources they're being cited from.

A baseline is 1 run. Engines aren't deterministic, so a re-run can move by an answer or 2 without anything having changed. A real move is one that holds across 2 runs, done the same way, at least a week apart.

What a Visibility Diagnostic delivers

The Visibility Diagnostic is the paid version of this, done for your business over 2 weeks. It's $1,500 and the fee credits toward a retainer if you continue.

You get:

• The baseline run above, logged and screenshotted
• A ranking check on the same queries in classic Google search, so you can see where rankings and citations disagree
• A citation-gap map: for each query where you weren't named, which sources the engine used for the businesses it did name, and whether those sources are open to you
• A fix list in the order we'd do it, with the ones you can do yourself marked

You don't get a report that tells you to "improve your content." Every item on the fix list is a specific page, listing or profile with a specific change.

What we don't claim

• We don't claim to control the engines. No one can guarantee an AI answer will name a business. We can change what the engines have available to verify about you, and we can measure whether the answer changed.
• We don't claim 1 run is a trend. A baseline is a dated snapshot. A before-and-after needs the same queries, the same engines, the same logged-out method and at least 2 runs on the "after" side.
• We don't claim impressions. The number FirstCited reports is consults booked. Being named is the leading indicator, not the result.
• We don't report what a visitor can't see. If a finding only shows up in source code, it doesn't go in the report as a visibility problem.
• We don't take clients the engines already name first. If your baseline shows you at the top on most queries, we'll tell you and the diagnostic ends there.

Common questions

What does FirstCited measure?

FirstCited measures whether AI answer engines - ChatGPT, Gemini, Perplexity and Google's AI answers - name a business when a buyer asks a buying question, and whether they link to a page the business controls when they do. Every run is dated, logged and screenshotted.

How is an AI visibility baseline run?

A short set of buyer queries, agreed with you first, is run through each engine logged out, with a new thread for every query, all in 1 sitting. For each answer we log the date, the engine, the businesses named, the order they're named in and every URL the engine links.

Why run queries logged out and in a new thread each time?

A signed-in engine personalizes on your history, and a question inside an existing thread is answered with that thread's context. Logging out and starting a new thread for each query gets the answer a first-time buyer would get.

What does a FirstCited Visibility Diagnostic include?

A logged baseline run, a classic Google ranking check on the same queries, a citation-gap map showing which sources the engines used for the businesses they did name, and a fix list in the order we'd do it. It's $1,500 over 2 weeks and the fee credits toward a retainer.

Can FirstCited guarantee an AI will name my business?

No, and nobody can. We change what the engines can find and verify about you, then re-run the same queries the same way and show you whether the answer changed.