Is SEO dead for law firms? What still decides the AI recommendation

Is SEO dead for law firms?

SEO is not dead for law firms. Google’s Search Central guidance on AI features says there are no additional requirements to appear in AI Overviews or AI Mode, and no special optimizations necessary, and that the same foundational search practices apply. It also says explicitly that a firm does not need to create new machine readable files, AI text files, or special schema.org structured data to appear in these features. Eligibility runs through the ordinary index.

That first-party statement covers Google’s surfaces, which is where the volume still is. It does not cover ChatGPT, Perplexity or Claude, each of which retrieves on its own terms. But the practical consequence holds across engines: a page that cannot be crawled, indexed or quoted is not a candidate anywhere, and almost every law firm site with an AI visibility problem also has an indexation or entity problem underneath it.

The scale of the shift is often overstated in one direction and understated in the other. Martindale-Avvo’s April 2026 analysis put AI tools at under 5% of total law firm website traffic, and Great Jakes measured AI referrals at 0.47% of sessions across its law firm client portfolio. A firm that dismantles the work producing the other 95% because of the 5% has misread the arithmetic.

What did AI search kill, and what did it leave alone?

AI search did not kill SEO as a discipline. It changed which SEO tasks pay. The tasks that build a retrievable, unambiguous record of the firm became more load-bearing, because an engine assembling an answer needs to resolve the firm as one entity and lift a sentence about it. The tasks built around winning a click from a blue link lost most of their value, because the click they were written to capture increasingly does not happen.

The table below is Firm AEO’s task-level read of that shift. It is a judgement about where budget goes, not a measured ranking, and the middle column is the part a firm can verify for itself in Google Search Console within a week.

Does ranking number one still get a law firm cited?

Not reliably, and it is not required either. Ahrefs’ March 2026 study of AI Overview citations found roughly 38% of cited pages ranking in Google’s top ten for the query, down from 76% in mid-2025. Of the rest, about 31% ranked between positions 11 and 100, and about 31% did not rank in the top 100 at all. Ranking is neither sufficient for citation nor necessary for it.

Both halves of that split matter to a law firm, and most coverage only quotes the first. The falling share says a firm holding position one for its main practice term cannot assume the Overview above it is quoting that page. The two thirds arriving from page two and beyond says a firm that gave up on a term because it sits at position 34 may still be quotable, provided the page answers the question in liftable sentences. That is a cheaper problem to fix than a ranking problem.

How much of an AI recommendation is still decided by SEO?

Nobody has published a defensible number, and the figure most often quoted does not survive a check. The claim that roughly 70% of what decides an AI recommendation is still traditional SEO circulates widely in legal marketing writing. Firm AEO went to the pages that carry it and could not find the figure, an attribution or a methodology at the source. It is not used here, and a firm should treat any agency that quotes it as owing an explanation.

What can be stated is narrower and more useful. Google says eligibility for its AI features runs through the standard index, which makes technical SEO a precondition on Google’s surfaces rather than a percentage. Off Google, the dependence varies by engine: Martindale-Avvo’s April 2026 comparison found ChatGPT matching Google’s top ten legal results under 25% of the time, Gemini near 50%, and Perplexity and Claude near 75%. A firm that is invisible in Google is invisible in Perplexity for the same reasons. The same firm can be invisible in ChatGPT for entirely different ones.

The practical reading is that traditional SEO is a floor rather than a share. It sets the ceiling on what the Google-adjacent engines can do with the firm, and it does roughly nothing on its own for the engines that read directory profiles, review platforms and forums instead of the firm’s own site.

If AI sends under one percent of sessions, why change anything now?

Because of the direction and the lag, not the current share. Previsible’s data, summarized by OpenLens, shows legal growing from 0.37% to 0.86% of total sessions between January and May 2025, with legal AI penetration up 11.9 times year over year across a sample of 1.96 million LLM sessions. Previsible’s July 6, 2026 report put ChatGPT at 92.4% of standalone large-language-model referrals across 166 properties and 6.77 million AI sessions. Legal is a fast-growing share of a small base.

The lag is the operative part. Fixing entity data, correcting directory profiles and rewriting a practice page into quotable form are all slow to register, because an engine has to re-crawl and re-corroborate before anything changes in an answer. A firm that waits until AI referrals are worth chasing starts a multi-quarter project at the moment it needed the result. Typical timelines run one to three quarters before appearance rate moves, which is an observed pattern in this work rather than a commitment.

There is a second reason that has nothing to do with referrals. The AI answer is increasingly the first description of the firm a prospective client reads, whether or not it produces a session. A firm can lose a matter to a sentence it never saw, in a session that never touched its analytics.

What should a law firm keep, cut and add?

The budget shift is smaller than the rhetoric suggests. Most of a competent law firm SEO program survives intact, one line item shrinks, and two get added. The list below is an allocation guide. The measures that tell a firm whether any of it worked are set out in Firm AEO’s post on rankings holding while traffic fell, which covers appearance rate, cited-source share and matters opened.

• •Keep technical SEO and indexation at full weight. It is now a precondition for citation on Google’s surfaces, not a hygiene task.
• •Keep and expand sub-practice pages, written so each one answers a specific question in sentences that stand alone when quoted.
• •Keep local SEO and the Google Business Profile. The local pack still appears on the queries that produce consultations.
• •Cut content produced to hit a monthly page count. Google’s scaled content abuse policy treats pages generated mainly to move rankings as spam, regardless of who or what wrote them.
• •Cut the reporting weight given to position tracking on head terms, without cutting the work that earns those positions.
• •Add directory and profile maintenance as a standing line item, since directories supply most of the citations engines return for legal queries.
• •Add a monthly appearance-rate test across a fixed question set, run logged out, which is the only number that shows whether any of the above reached an answer.