What to track when clicks decline but calls rise
If your website analytics show fewer clicks but your phone rings more and intake forms keep arriving, the signal to trust is intake data, not click counts. AI search tools like ChatGPT, Gemini, Perplexity, and Google AI Overviews often answer a person's question directly, so the person never clicks through to your site before calling. Measuring whether AI search brings new clients means tracking calls, form fills, and how new clients say they found you, alongside traffic.
Why traditional traffic numbers understate AI referrals
Website analytics tools were built to count clicks from search results pages, and they do that job well for traditional search. AI answer engines change the path: a person asks "who handles child custody cases near me" and gets a direct answer with your firm's name, phone number, or a summary of your services, without ever landing on your website. That interaction never shows up as a website session, even though it produced a client. Firms that only watch website traffic will conclude AI search isn't working when it may be working quite well.
This gap matters because a prospective client who gets your name and number from an AI-generated answer often calls directly instead of browsing your site first. That call looks, in your phone system, identical to a referral from a friend or a billboard. Without asking, you cannot tell the difference. The traffic report stays flat or declines while the client roster grows, and the two numbers seem to contradict each other when they are actually describing different parts of the same client journey.
Asking intake how new clients found you
The single most reliable way to measure AI search impact is asking every new client, at intake, how they found your firm. This means training whoever answers the phone or processes a new client form to ask a simple, consistent question and record the answer in the same place every time. Over weeks, patterns emerge that no traffic tool can show, because the client is the only source of truth for what happened before they picked up the phone.
The question needs to be specific enough to catch AI search without leading the client toward an answer. "How did you hear about us?" often produces vague replies like "online" or "I searched." A better follow-up is "Did you use a search engine, ask an AI tool like ChatGPT, get a referral, or see us somewhere else?" Naming AI tools directly in the question gives clients permission to say "I asked ChatGPT" instead of defaulting to "I googled you," which is what many people say regardless of which tool they actually used. Consistency in asking and logging this question, across every intake call for a full month, is what turns anecdotes into a usable pattern.
Watching branded search and direct inquiries
A rise in people searching your firm's name directly, or typing your web address instead of searching a general term, often means an AI tool mentioned your firm by name and the person went looking for you afterward. This behavior, sometimes called a branded search lift, shows up in the same analytics tools you already use, even when it doesn't credit AI search as the source. Watching this trend over time gives you a second signal that works alongside intake questions.
To watch for this, check your website analytics for search terms that include your firm's name, and compare that volume month over month. Also watch for an increase in people who type your web address directly into a browser, or who arrive with no referring source listed at all, sometimes labeled "direct" traffic. A jump in either of these, especially if it lines up with new clients mentioning an AI tool at intake, suggests the two are connected. Neither metric alone proves AI search drove the visit, but together with intake answers, they build a picture traffic totals cannot provide on their own.
Building a simple monthly check
A monthly check that takes an hour can tell you more about AI search impact than a dashboard full of traffic charts. The check has three parts: reviewing what intake recorded for "how did you find us" answers that month, comparing branded and direct search numbers to the prior month, and reading through any new client conversations for mentions of AI tools by name. Doing this consistently, on the same date each month, turns scattered observations into a trend you can act on.
Start by pulling the intake log and counting how many new clients mentioned an AI tool, a search engine, or a referral source. Next, open your analytics and note branded search volume and direct traffic for the month, writing the numbers down somewhere you'll compare them to next month rather than trusting memory. Finally, skim any recorded consultations or intake notes for phrases like "I asked ChatGPT" or "an AI thing told me about you." None of this requires new software. It requires a habit, a notebook or spreadsheet, and the discipline to ask the same question the same way every time a new client calls.
Once you have two or three months of this check, you'll notice whether AI-related mentions are rising, flat, or falling, and whether that lines up with branded search movement. That trend, not any single month's number, tells you whether your visibility in AI-generated answers is actually converting into signed clients. A firm that sees AI mentions rising alongside branded search and steady or growing intake has real evidence that AI search is working for them, independent of what the raw traffic report says.
What to ask a marketer before you hire them for this
Before hiring anyone to help your firm show up in AI search results, ask them how they would measure success without relying only on website traffic. A marketer who understands AI search should be able to describe an intake-tracking approach, mention branded search as a signal, and explain how a prospective client can get your firm's name from an AI tool without ever visiting your site. If they can only talk about clicks, rankings, or traffic dashboards, they are describing the old search landscape, not the one your prospective clients are actually using.
Ask them directly: "If a person asks ChatGPT who handles cases like mine and gets your firm's name with no click to our website, how would we know that happened, and how would you show me it's working?" Ask what they would want your intake team to record starting the first week of the engagement. Ask how they would separate a branded search increase caused by AI mentions from one caused by a billboard or a referral spike. A marketer with real understanding of AI search will have specific, practical answers to each of these questions, because they will have already thought through the gap between what analytics shows and what actually brings a client through your door.