Can a law firm website be cited in Google AI Overviews? A reproducible search test

Yes in principle: any indexed, readable page can be used as a supporting link in a Google AI Overview, and there is no category of site that is excluded in advance. Whether a particular page is selected is a different question with no published rule, and nobody outside Google can measure it in general. As of 29 September 2026 the number of queries we have tested for an AI Overview linking aiknowsus.com is zero, so the count is 0 out of 0 and no share appears on this page. What appears is the exact query design, the observation record fields, the arithmetic, and the nearest measured fractions we hold, which come from ChatGPT, Claude and Perplexity in two named sectors rather than from Overviews.

One thing has to be said before the method. aiknowsus.com is not a law firm. It is a measurement product. Counting law firm queries against our domain would be a test built to return zero, and calling that a finding about law firms would be worse than publishing nothing. So the count above is about our own domain, and the query design below is for a law firm to run on its own.

The answer, first

Four statements, each exact.

• Eligibility. A page that is indexed and readable in the form a crawler receives is in the pool of pages that can be used. That is eligibility, and it is not an expectation.
• Selection. How the displayed links are chosen from the pages considered is not published, so no method can promise you one.
• Queries we have captured: 0. As of 29 September 2026 we hold no AI Overviews capture of any kind, so there is no share of queries linking our domain on this page.
• The nearest measured fraction on our domain: 0 of 6. Perplexity, September 2026, six blind questions in AI visibility and generative engine optimisation tooling. Not cited or recommended in any of the six answers. That is an assistant, not an Overview, and we are not offering it as a substitute.

The nearest record we hold that a specific page can move from absent to cited is a single one: on 6 August 2026 ChatGPT cited clawlaw.in/blog/how-to-check-a-companys-court-cases-in-india as a source for a vendor due diligence question, fourteen days after that page was published, in a zone where the same question set had named the company nowhere at the 27 July 2026 baseline. One page, one engine, two dates. Fourteen days is a single observation, not a median and not an average, and anybody quoting you a typical time to first citation should be asked for their denominator.

How it was measured

The test, in eleven steps, none of which needs a paid tool. What has not been run is named in the next section.

One. Build the query set from named markets and named practice areas. Seven markets: Mumbai, Delhi, Bengaluru, Hyderabad, Chennai, Pune and Ahmedabad. Six practice areas: company and commercial litigation, insolvency under the Insolvency and Bankruptcy Code, employment and labour, intellectual property, real estate and RERA, and direct and indirect tax. Write one to three queries per pair in the words a client types, which are longer and more specific than the words a keyword tool favours. Save the file with a version number and a date.

Two. Split the file three ways and keep three denominators: informational queries about how a process works, comparison queries about how to choose a firm or an adviser, and hiring queries naming a place. The three behave differently and merging them produces a number about none of them.

Three. Put your firm's name in none of the queries. On 27 July 2026, in the clawlaw.in programme, two runs of the same 78 questions happened on the same day. The first used a wrapper that named the brand and came back ranking it first on almost every question. The second named nothing and put the company second by breadth and absent altogether from the litigation due diligence questions it most wanted to win. The first was discarded, because the only thing it had measured was our own prompt. ChatGPT, 27 July 2026.

Four. Use a clean browser profile, signed out, with no extensions, and record the browser and its version.

Five. Set the location explicitly and record the method. A location set in search settings, a VPN, and a device physically in the city are three different observations. Write which one you used on every row.

Six. Capture desktop and mobile as separate rows, because they are two observations and merging them makes a trigger count meaningless.

Seven. Let the page settle before capturing, so an Overview that loads a second late is not recorded as absent.

Eight. Save three artefacts per query: a full page screenshot, the Overview text pasted as text, and every supporting URL copied exactly.

Nine. Fill one observation record per query with these sixteen fields: query text, file version, query type, market, practice area, location method, city, device, signed in state, timestamp, Overview present yes or no, number of links shown, own domain linked yes or no, firm named in the answer text yes or no, competitor domains linked, and the screenshot filename.

Ten. Repeat each query three times in the same pass where you can, treating each as its own observation, because an Overview can differ between two searches minutes apart.

Eleven. Label twice and rerun on a fixed interval, with a second person labelling 20 rows blind and the agreement rate published, and the query file unchanged between passes.

Run three Search Console checks alongside the capture. Whether each page you expect to be used is indexed at all, using URL Inspection. What Google actually received, by reading the fetched page rather than your published page. And which queries already bring impressions without clicks, which is the best source for the query file. Search Console will not confirm that a particular Overview used your page, so do not build a report on a breakdown you have not verified exists on the day you read this.

The arithmetic, printed so any figure we publish can be checked. 42 pairs at one query each, across two devices, is 84 observations. If 47 of those 84 show an Overview, the trigger line is 47 of 84 in the named cities on the stated dates, and every Overview based metric after that uses 47 as its denominator. If 3 of the 47 link your domain, that is 3 of 47, written that way. Reporting 3 of 84 merges two gates and explains neither. This is arithmetic, not an observation.

What the numbers were

Queries captured for AI Overviews: 0. Queries where an Overview linked aiknowsus.com: 0 out of 0. Share: not calculable. Those cells are empty on purpose, and no figure borrowed from another study will be placed in them. When the capture runs, this section will carry the query count, the Overview count, the linked count, the cities, the location method and the dates in one sentence.

The measured fractions we hold, each with engine, date, sector and denominator.

• 0 of 6, aiknowsus.com. Perplexity, September 2026, AI visibility tooling. Not cited or recommended in any of six blind answers.
• 161 across 24 batches and 72 conversations, aiknowsus.com. Perplexity, September 2026. Statements in the engines' own self audits recording that we were not cited.
• 1 of 18, clawlaw.in. ChatGPT, 18 August 2026, Indian legal research software. Top source on exactly one of eighteen blind commercial questions.
• 0 of 6 for third party sources, clawlaw.in. Claude, 17 September 2026. No third party review or directory source reached any of six answers, and one well known review site was dropped from the raw results because its list was of American products.
• 28, 21 and 18 out of 78, with 0 of 78 searched. ChatGPT, 27 July 2026. The blind breadth order was ProVakil, CLAW and Legistify, from a run in which the engine confirmed it had performed no live web search at all.
• 1 page, 14 days, 1 engine. ChatGPT cited a named clawlaw.in URL on 6 August 2026, fourteen days after publication. One observation, not a median.

None of those are Overviews. The systems select sources differently and a result from one is not evidence about another, which is why the empty cell above stays empty.

Two dated records explain classes of absence that an Overviews audit will otherwise record as bad luck. On 6 August 2026 the clawlaw.in pricing page rendered its prices only after scripts ran, so a crawler received a page with no prices at all and the prices the assistants quoted came from an app store listing instead. On 17 September 2026 Claude ranked two enterprise vendors above clawlaw.in specifically because they publish explicit court and tribunal coverage lists, and noted that its ordering reflected price transparency and source authority rather than product quality. The first is a technical failure and takes five minutes to find. The second is a publishing failure and takes an afternoon to fix.

What this cannot tell you

Eight limits.

• There is no Overviews data behind this page. Zero queries captured, stated again because it is the sentence most likely to be skipped.
• A snapshot is not a rate. 47 of 84 in seven cities on three days would be a fact about those queries, cities and days, and nothing wider.
• Triggering is unstable, so a fall between passes is not necessarily a loss of visibility.
• Location method changes the answer, and the size of that effect is not published, so two audits done differently are not comparable.
• The displayed links are not the full source set. On 17 September 2026 Claude admitted using two arguments from clawlaw.in pages and dropping the attribution in both places. A page can be used and not shown, and no audit sees that.
• Search Console cannot confirm that an Overview used your page. Do not build a report on a breakdown you have not verified exists today.
• Our measured figures come from two sectors that are not law firms, and each line names its sector for that reason.
• No method promises a position in an Overview. Nobody can sell you that, and a supplier who offers it is telling you something about themselves.

The limit that matters most to a law firm in India is not a measurement limit. What a firm or an advocate may publish about itself is governed by the Bar Council of India's rules on advertising and solicitation, and those rules decide what belongs on your pages before any visibility method does. Read the current rules yourself and take advice where you are unsure. Nothing on this page is legal advice.

Sources and change log

Every figure above traces to one of two recorded bodies of work. The clawlaw.in programme from July 2026, recorded in GEO_BASELINE_RESULTS_2026-07-27.md, GEO_GAP_ANALYSIS_2026-08-18.md and claude_response_17_09_audit.md. And the aiknowsus.com audit of September 2026, 24 batches and 72 conversations, captured outside the site repository. Counts were produced by a script over those files rather than from memory. Of the observations used here, the 6 August 2026 citation of a named URL can be verified from outside today; the counts over our capture folders cannot, until the captures are published.

We restate no Google documentation on this page. Read Google's current help material on AI Overviews, the Google Search Central documentation, and Google Business Profile's help pages yourself, and record the URL and the date beside anything you take from them, because help pages are edited and an undated paraphrase becomes a false claim the page cannot detect. We restate no competitor's prices anywhere on this site either: open the vendor's own pricing page and record price, date and URL together in one row. On 17 September 2026 Claude checked two of our comparison pages whose competitor prices were correct and undated, and used official sources instead.

Version: 29 September 2026, first publication. Updates will be triggered by our first Overviews capture, in which case 0 out of 0 becomes a real fraction with a query file version, a location method and dates; by any correction to a figure above; and by publication of the capture files. Corrections are noted here rather than made quietly.

Common questions

Is a law firm website allowed to appear in an AI Overview?

There is no published exclusion for any category of site. If a page is indexed and its content is readable in the form a crawler receives, it is in the pool that can be used. Selection is separate and unpublished, so the honest answer to whether yours will be selected is that nobody can tell you in advance.

Which pages on a firm's site are most likely to be usable?

The ones that answer a single question a client actually asks, with the answer in the first paragraph and a checkable fact underneath. The nearest measured evidence for that shape is from another sector: on 17 September 2026 Claude ranked two vendors above clawlaw.in because they publish explicit court and tribunal coverage lists, and on the same date placed official portals above every commercial page for questions about using those portals. A firm's equivalent is naming the courts and tribunals it appears before and the practice areas it takes instructions in.

We are indexed and still never appear. What do we check?

Read the fetched version of the page in URL Inspection and confirm that the facts a client needs are in what Google received, rather than added by a script after load. On 6 August 2026 that single check explained why assistants were quoting an app store listing for clawlaw.in prices instead of the company's own pricing page.

Should we publish client results to get cited?

Only what the Bar Council of India's rules permit, and only with a date and a source a reader could check. One unverifiable claim can cost you the verifiable ones: on 6 August 2026 a headline figure on clawlaw.in could not be true, and when Claude checked it against public numbers it advised a buyer against the product, after which the company's accurate claims stopped counting for that answer.

How many queries make a useful first capture?

Twenty, in the two or three practice areas and the one or two markets you actually want more work in. Forty two pairs is a complete study and a poor first week. Whatever you run, print the denominator beside every result.

Can we pay for a place in an AI Overview?

No, and any offer of one should be treated as a reason to stop the conversation. We sell measurement and page writing, and we still cannot promise a position in an answer.

What to do first

Write 20 queries in a client's words across your two most important practice areas and the city you actually take instructions in, with your firm's name in none of them, and save the file with today's date and the location method you will use. Capture all 20 today, signed out, on desktop and mobile as separate rows, recording Overview present, links shown, own domain linked and firm named. Then run URL Inspection on your three most important pages and read what Google actually received. That is one afternoon, and it gives you a dated baseline plus, most of the time, the explanation for your absences.