--- title: "Measuring AI Search Visibility for Law Firms: What to Track" description: "Measuring AI search visibility for law firms starts where the ranking report stops. What to track, what to drop, and the one number nobody can attribute." url: "https://carlosarias.com/blog/law-firm-marketing/ai-search-measurement-law-firm-marketing" --- [Law Firm Marketing](/blog/categories/law-firm-marketing) # Measuring AI Search Visibility for Law Firms: Your Report Still Leads With Rankings Measuring AI search visibility for law firms starts where the ranking report stops. What to track, what to drop, and the one number nobody can attribute. [Carlos Arias](/blog/authors/carlos-arias) · September 20, 2026 · 7 min read ![An ink-brush figure faces a small open circle beneath a large pale ring, with a red seal at lower right.](/_astro/cover.BFz3bjrX_2wEiGN.webp) *An ink-brush figure faces a small open circle beneath a large pale ring, with a red seal at lower right. AI-generated illustration by Carlos Arias .* Measuring AI search visibility for law firms starts with admitting the ranking report has stopped predicting anything. Position three used to mean traffic. In 2026 it means you were eligible for a click that an AI summary may have already absorbed on the way past you. If the monthly PDF from your agency still opens with average position and a green arrow, you are reading a proxy for a proxy. There is a real stack underneath this, and it has four pieces. Google shipped two of them this year. The third you build yourself. The fourth cannot be attributed at all, not by you and not by anyone selling you a dashboard. This piece walks all four. ## Rankings and citations came apart, and we can date it This is not a vibe. Ahrefs matched 863,000 keywords against 4 million AI Overview URLs and found that only 38% of cited pages ranked in the top 10 for the query that produced the citation. Seven months earlier the same measurement read 76%. It halved. So more than six in ten AI Overview citations now go to pages that are not on page one for the query at all. Your rank tracker cannot see them. It is watching a leaderboard the citation engine stopped consulting. Then there is what happens to the clicks that remain. Pew Research installed browsing trackers on more than 900 consenting US adults and logged 68,879 Google queries across March 2025. When an AI summary appeared, users clicked a traditional search result in 8% of searches, against 15% when no summary appeared. The summary ate the click. Clicks on the links inside the summary were rarer still, about 1% of visits. Zero-click behavior on Google reached 68% of US queries in early 2026. Put those together for a personal injury firm. You can hold position two on a city-plus-practice term and watch impressions rise, and still sign fewer consults than you did last year. Every number on the report will look fine. The report is measuring a door that fewer people walk through. ## Measuring AI search visibility for law firms: the three sources that replace it None of these is a rank tracker with an AI skin on it. They measure different things. They disagree with each other on purpose. ### 1. The Search Console generative AI report, read as impressions only Google added a dedicated generative AI performance report to Search Console and rolled it out to every site worldwide on August 31, 2026. It counts an impression whenever one of your URLs appears inside AI Overviews or AI Mode, and you can slice it by page, country, device and date. Read Google’s own documentation on what it leaves out. There are no clicks in it and no CTR. No query data either. You learn that a page surfaced. You do not learn what was asked, or whether anyone moved. That is still the single most useful new number a firm has. It tells you which practice pages are citation-eligible and which are invisible to the layer that now answers first. Treat a rising impression count as proof of eligibility, never as proof of demand. One asterisk sits on the history. A logging error deflated impressions in the generative AI report for data from August 13 to August 17, 2026, and Google recorded it in the Search Console data anomalies list before restoring the missing days on August 21. Search Engine Land covered the drop while site owners were still reading it as a collapse in visibility. Any month-over-month comparison spanning that window carries the asterisk. ### 2. GA4’s AI Assistant channel On May 13, 2026, Google Analytics added AI Assistant as a default channel group, so referred sessions from recognized assistants get a medium of ai-assistant instead of disappearing into generic referral. Search Engine Journal covered the rollout. No configuration required, which is rare enough in GA4 to be worth saying out loud. Two cautions before you put it on a slide. Channel group changes apply going forward, so do not expect your 2025 history to reclassify itself and hand you a trend line. And the channel only captures assistants that pass a referrer Google recognizes. It is a floor, not a count. ### 3. A prompt panel you run yourself There is no product that does this honestly yet, so build it. Write 30 to 50 prompts in the words a person in trouble actually uses, not keywords. Not “boca raton probate attorney” but “my dad died in Florida without a will and my brother has the house keys.” Then run them against the assistants your clients use, on a fixed cadence, and log three states per prompt: cited with a link, mentioned without a link, absent. Mention-without-link matters more than people assume, because a named firm gets typed into Google later. - Run each prompt five times, not once. These systems are non-deterministic. A single run tells you almost nothing, and a firm that celebrates one lucky citation is measuring noise. - Record the competitor set too. The useful finding is usually which three firms keep appearing and what their pages have that yours does not. - Log the source URLs the answer cites. Half the time it is a directory or a bar association page, which tells you the work is a profile problem rather than a content problem. It is tedious. No point pretending otherwise. It is also the only part of the stack that tells you what the machine says about your firm to a person deciding who to call. ## Stop tracking these Average position across a keyword basket. It aggregates unlike queries into a number that moves for reasons nobody can act on. Total organic sessions as a headline. Sessions can fall while qualified inquiries rise, and a report that cannot tell those two stories apart is not a report. “Keywords in top 3.” The count was always a proxy for demand capture, and the Ahrefs citation data shows the proxy broke. ## The number that was always the real one The real number is signed matters by source. Not leads. Not calls, and not the number of times the phone rang. Signed. A ranking chart takes four seconds to produce. A defensible count of matters signed from local search last quarter takes a week of digging, and in most firms the answer is not recorded anywhere at all. That gap is the actual reporting failure, and it predates AI by a decade. The fix is unglamorous. A source field on the intake record that a human is required to fill, reconciled monthly against the channels you actually paid for. Pair it with one operational metric, median response time to a new inquiry. A firm that is cited by every assistant in South Florida and answers at 9am the next morning has bought expensive lead leakage. We have written about where law firm leads actually die, and the pattern has not changed because the search layer changed. If your practice pages are not getting cited at all, the input problem is usually language rather than markup. That is a separate piece of work, and it starts in your own intake transcripts rather than a keyword tool. The ordering of the whole search program, local first and AI eligibility after, is laid out in our 2026 law firm SEO guide. ## The constraint nobody can engineer around Here is the honest part, and any vendor who skips it is selling you something. AI-assisted referrals cannot be fully attributed today. Not by anyone. Some assistant interfaces strip the referrer, so those sessions land in Direct and look like someone typing your URL from memory. Worse for measurement, the highest-intent path barely touches your analytics at all. A person asks an assistant who handles wrongful termination in Palm Beach County, reads three firm names and closes the tab. Two days later they search the firm by name. That arrives as branded organic. It was an AI referral. Nothing in your stack will ever say so. The workaround is to track branded search volume as a shadow metric alongside the prompt panel. If your citation rate in the panel climbs and branded impressions climb behind it with a lag, you have a directional signal. It is an inference, not a measurement. Label it that way in the report. Modeled attribution presented as fact is how agencies keep retainers alive past their usefulness. A report with an honest hole in it beats a clean chart built on an assumption. Name the hole. Measure everything around it. ## If this is your report Pull the Search Console generative AI report for the last 90 days and look at which practice pages have zero impressions in AI features. That list, by itself, will tell you more than the last six monthly decks combined. It costs nothing and takes about ten minutes. If what you find there does not match what your agency has been reporting, that is worth a conversation. Write us at hi@carlosarias.com. Tags [#Law Firm SEO](/tags/law-firm-seo/)[#AI Search](/tags/ai-search/) #Analytics [#Law Firm Marketing](/tags/law-firm-marketing/)[#Local SEO](/tags/localseo/) Share Written by [Carlos Arias](/blog/authors/carlos-arias) Marketing Engineer for law firms. I combine digital marketing, software, data, automation and AI to improve the whole system — from first click to signed case. On this page - Rankings and citations came apart, and we can date it - Measuring AI search visibility for law firms: the three sources that replace it - 1. The Search Console generative AI report, read as impressions only - 2. GA4’s AI Assistant channel - 3. A prompt panel you run yourself - Stop tracking these - The number that was always the real one - The constraint nobody can engineer around - If this is your report ## Continue reading [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 11, 2026 [### SEO for Law Firms in 2026: Local First, Then Practice Pages, Then AI Citations](/blog/law-firm-marketing/seo-for-law-firms-2026/) Law firm SEO in 2026 starts with the map pack, then practice-area pages that convert, then eligibility for AI citations. Timelines, intake, and what agencies get wrong — with no ranking guarantees. Carlos Arias · 14 min [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 18, 2026 [### Law Firm Website AI Visibility: It's an Architecture Problem](/blog/law-firm-marketing/machine-first-architecture-law-firm-website/) Law firm website AI visibility is an architecture problem, not a content problem. The four pillars, what breaks on a practice-area page, and the fix. Carlos Arias · 8 min [Law Firm Marketing](/blog/categories/law-firm-marketing) · September 11, 2026 [### How Law Firm Marketing Actually Works: The Systems Guide (2026)](/blog/law-firm-marketing/law-firm-marketing-systems-guide-2026/) Law firm marketing is not posts and ads — it is the system from search to signed case. SEO, local, paid, website, intake and measurement, with the places firms lose cases they already paid for. Carlos Arias · 18 min ## Stay in the loop. One email when it’s worth it — new posts and updates, no spam. Thanks — check your inbox to confirm. Free. Unsubscribe in one click.