Something quiet happened to legal search in Australia, and the pattern is not what most law firms assume.
In September 2026 we ran 43 Australian Google searches and recorded every one that returned an AI Overview. Then we sorted the results by intent. The split was almost total, and it changes what a law firm should be doing with its content.
Every informational query in the test returned an AI Overview. Almost no commercial query did. For law firms specifically, the divide was absolute: eight out of eight informational legal searches returned one, and not one of the thirteen commercial or local legal searches did.
That finding is the point of this article. The AI layer is answering legal research, not legal hiring, and those two jobs now need completely different work. This piece breaks down what we measured, what it means for a firm’s content plan, and how we at 21 Webs structure law firm SEO around it.
Key Takeaways
- 21 Webs tested 43 Australian Google searches in September 2026, desktop, using live SERP data, and counted an AI Overview as present when the results page returned one.
- Informational queries returned an AI Overview 19 times out of 19. Commercial or local queries returned one 1 time out of 24.
- Within legal, 8 of 8 informational searches returned an AI Overview, and 0 of 13 commercial or local legal searches did.
- All 13 commercial legal searches returned a local pack instead, and a local pack is territory a law firm can still compete in.
- The AI layer is eating legal research, not legal hiring.
- The sample is 43 queries from a single date, so it cannot rank industries against each other. The intent pattern is the finding, not the vertical split.
Intent Decides Whether an AI Overview Appears, Not the Industry
Here is exactly what we did. We ran 43 Australian Google searches from a desktop environment using live SERP data, and recorded whether each results page returned an AI Overview. Every query was run once, from one location, on one date in September 2026. Then we grouped the results two ways: by the intent behind the query, and by the industry it belonged to.
Sorting by intent produced the sharpest result in the whole dataset. Informational queries returned an AI Overview 19 times out of 19. Commercial or local queries returned one 1 time out of 24, a trigger rate of 4.2 per cent.
Sorting by industry produced a much messier picture. Legal ran 21 queries and returned an AI Overview on 8 of them, a rate of 38 per cent. We ran smaller sets in three other industries as well, and we are deliberately not publishing those figures side by side. A handful of queries per industry cannot carry a comparison, and a comparison the sample cannot support is exactly the kind of number that gets quoted back at you out of context.
It is tempting to read an industry figure as a ranking and conclude that one vertical is more exposed to AI search than another. That is the reading we want to head off. Twenty-one queries is not a sample that can carry a ranking claim, and a smaller set certainly is not. The honest reading is that the vertical numbers move around and the intent numbers do not.
Now the caveats, because they matter more than usual here. This is 43 queries. It is one snapshot, from one date, on desktop, with one location per query. It is not a national census of Australian search, and it should not be presented as one. The per-industry samples in particular are far too small to rank verticals against each other, and their confidence intervals overlap. What the data supports is a pattern, and it is a pattern strong enough to act on. What it does not support is a league table.
That distinction is the reason we are publishing the intent split rather than a vertical ranking. A ranking would be a claim the sample cannot carry. The intent split is a pattern a firm can build a content plan on.
Legal Informational Queries Are Being Answered. Legal Hiring Queries Are Not.
We ran 21 legal searches to get a clearer read on one industry, because legal is the vertical we chose to test in depth, and the one where the split matters most to the people searching.
Those 21 queries covered seven practice areas: conveyancing, family law, personal injury and compensation, immigration, criminal law, wills and estates, and employment. They were spread across four capital cities, Sydney, Melbourne, Brisbane and Perth, alongside national phrasing that did not name a location.
The split inside legal was absolute.
Informational legal queries returned an AI Overview 8 times out of 8. Every “what is”, “how does” and “how much” query came back with a generated answer sitting above the results.
Commercial or local legal queries returned an AI Overview 0 times out of 13. Not once, across thirteen attempts. Every one of those results pages returned a local pack instead, the map-and-listings block that sits below the ads.
The commercial queries that returned no AI Overview included “conveyancing lawyer Sydney”, “family lawyer Melbourne”, “personal injury lawyer Sydney”, “criminal lawyer Melbourne”, “lawyer near me” and “best lawyer australia”.
Read that list again, because it is the finding. Those are the queries a law firm’s service pages are built to win, and in our test the AI layer stayed out of all of them.
One exception worth naming, because it shows the pattern turns on intent rather than any mechanical rule. “Best real estate agent australia” did return an AI Overview. “Best lawyer australia” did not. Both are commercial queries opening with “best”, both point at a professional service, and one triggered a generated answer while the other did not. We are not going to pretend 43 queries explain why. What we can say is that the pattern held across 23 of the 24 commercial and local queries we ran, and a single exception does not change its shape. Anyone who tells you they can predict AI Overview presence query by query is overstating what anyone can currently measure.
What This Means for a Law Firm's Content Plan
The finding splits a law firm’s marketing into two jobs that used to be one.
The informational layer needs citation work, not traffic chasing. When a prospective client asks “how does child custody work in Victoria”, the answer now appears above your site and they may never arrive at it. Building a practice area page to rank for that query and then waiting for the visit is a losing plan, because the visit is exactly what the AI Overview removed. What you can still earn is the mention. A clear, well-structured, attributed answer to that question is one of the things an AI system has to choose between, and being chosen is worth more than a ranking nobody clicks.
The commercial layer still returns a local pack, and a local pack is winnable. Every one of the 13 commercial and local legal searches we ran returned a local pack and no AI Overview. That means the map block is still the battleground for “conveyancing lawyer Sydney”, and the factors that decide it are the ones a firm can control: a complete and accurate Google Business Profile, a steady flow of reviews, consistent name and address data across every directory, and service pages that make the location and the practice area unmistakable.
There is a strategic benefit hiding in that split. The two layers need different work, and most firms are doing neither well because they have not separated them. A firm that puts its content budget into citation-ready answers, and its local effort into the profile and the reviews, is working on the two things that still move rather than on a ranking that no longer produces the visit.
Entity Authority Is the Citation Prerequisite
Before AI will cite your law firm, it needs to know your law firm exists as a defined entity.
Entity authority means your firm appears consistently across:
- Google Business Profile with complete, accurate information, correct categories, and regular reviews
- Legal directories (Law Society registries, state bar association listings, legal-specific directories)
- Business directories (Australian Business Register, local council listings, industry associations)
- Social platforms with consistent name, address and branding
- Your own website with Organisation and LegalService schema markup that tells search engines and AI systems exactly what your firm is, where it operates, and what practice areas it covers
This is standard local SEO practice, but for law firms it carries additional weight because of the YMYL classification. Google and AI systems apply a higher trust threshold to legal content, and entity consistency is a foundational trust signal.
If your firm’s name is listed differently across three directories, your address is outdated on your Law Society profile, or your Google Business Profile has the wrong practice areas, AI systems cannot confidently attribute content to your firm. No attribution, no citation.
Think of entity authority as the answer to a question an AI system asks before it mentions anyone: is this a real, verifiable organisation, and is it the same organisation across every source I can see? A firm that appears as three slightly different names across three directories fails that test without ever being told. The failure is silent, it costs nothing to fix, and it never shows up in a ranking report.
Practice area classification deserves particular attention here. A firm that lists six practice areas where it genuinely handles six, and a firm that lists twenty because listing more feels safer, are not equivalent. The first is a defined entity. The second is an unclear one, and clarity is what gets cited.
There is a second reason this matters, and it connects directly to the intent split above. The commercial and local queries that still return a local pack are decided largely by entity signals. When someone searches “conveyancing lawyer Sydney”, the map block is Google’s best guess at which firms are real, nearby and credible. Entity consistency is how you enter that conversation.
E-E-A-T Is the Ranking and Citation Gate for Legal Content
Google classifies legal content as YMYL, meaning it can significantly impact a person’s health, financial stability, safety or wellbeing. YMYL content is held to a higher quality standard across four dimensions:
Experience: Does the content demonstrate firsthand experience with the topic? A page about family law written by a practising family lawyer carries more weight than a generic overview.
Expertise: Does the author have demonstrable qualifications? For legal content, this means law degrees, practising certificates and specialist accreditations.
Authoritativeness: Is the firm recognised by authoritative sources? Citations from legal publications, court listings, Law Society registries and government sites build authority.
Trustworthiness: Is the content accurate, transparent and compliant? For Australian law firms, this includes compliance with ASCR Rule 36 (no false, misleading or deceptive claims) and Australian Consumer Law.
AI systems use the same signals. When a system selects sources for a legal answer, it favours pages from entities with strong E-E-A-T signals. A practice area page with author attribution, schema markup, correct legal citations and links from authoritative legal sources will outperform an anonymous blog post on the same topic.
The “Experience” dimension is where most firm websites fall short, and it is the easiest one to fix. A page with no author is an anonymous page, and an anonymous page cannot demonstrate experience no matter how well it is written. Adding a named practitioner, their admission date or jurisdiction, and their area of practice turns an anonymous page into an attributed one. That single change does more for citation eligibility than any amount of rewriting.
Trustworthiness is the dimension where legal differs from every other vertical, and it is the one most likely to be underestimated. In most industries, an ambitious claim is a marketing risk. In legal, it is a compliance risk as well, because ASCR Rule 36 applies to everything published in connection with a practice. The good news, which the compliance section below covers, is that the compliant approach and the citable approach are the same approach.
We covered how AI search reshapes SEO in a recent analysis. For legal, the E-E-A-T bar is higher than for any other vertical.
Direct-Answer Headings, Schema and Comprehensive Guides Win Citations
The structural requirements for AI citation in legal are specific and measurable.
Direct-answer headings
AI systems parse content by heading. A heading that asks the question the user searched (“How does property settlement work in Victoria?”) followed by a direct, concise answer in the first paragraph is far more likely to be cited than a heading that describes the firm’s service (“Our Property Settlement Services”).
Write headings as questions your clients actually ask. Answer them directly in the first two sentences. Then expand.
This is a bigger change to legal content than it sounds. Most law firm pages are organised the way a firm is organised: by service. Clients do not search by service, they search by problem. Reorganising a page around the client’s question rather than the firm’s offering is the single highest-leverage structural change available, and it costs nothing except the decision to do it.
Schema markup
Legal content should carry:
- FAQPage schema on FAQ sections and practice area pages
- LegalService schema on practice area and service pages
- Organisation schema on the firm’s homepage and about page
- Attorney schema (or Person schema with legal credentials) on team/profile pages
- BreadcrumbList schema on every page for site hierarchy clarity
Schema does not guarantee citation. But it makes your content machine-readable, which is a prerequisite for being parsed by AI systems.
Comprehensive guides over thin pages
A single comprehensive guide on “Separation and Divorce in Australia”, covering grounds, process, property, children and costs with structured headings and FAQ schema, will outperform a set of thin pages built one per subtopic. AI systems favour substantial, authoritative pages that answer a cluster of related questions.
That preference runs against a common law firm instinct, which is to build a short page for every service and every variation of every service. The instinct comes from traditional SEO, where a page per keyword was the play. AI answers reward depth over fragmentation, because a comprehensive guide on one subject gives the system a complete, coherent source rather than five partial ones. Fewer, better pages is the direction.
There is a resourcing implication in that too. Consolidating several thin pages into one comprehensive guide is less content to maintain, less to update when legislation changes, and a stronger single asset for citation. Firms that worry about the rewrite usually find the total word count ends up lower rather than higher.
Verifiable Content Is What AI Rewards, and What ASCR Rule 36 Requires
Here is a compliance finding that also happens to be a citation finding.
ASCR Rule 36 requires that all advertising, marketing and promotion connected to a solicitor or law practice must not be false, misleading or deceptive, offensive, or prohibited by law. This applies to websites, SEO titles, meta descriptions, Google Ads, social media posts, directory listings and review responses.
The practical effect: every claim on your website must be verifiable. You cannot claim to be a “specialist” unless you hold formal specialist accreditation from your state Law Society. You cannot imply outcomes that are not typical. You cannot use testimonials in a way that creates unrealistic expectations.
AI systems reward the same discipline. Verifiable, factual, well-sourced content is what gets cited. Vague claims, superlatives and unattributed assertions are what gets ignored.
For law firms, ASCR Rule 36 compliance and citation readiness are the same checklist:
- Make factual, verifiable claims about your practice areas
- Attribute content to named, qualified authors
- Cite legislation and case law accurately
- Avoid guarantees, implied outcomes and unsubstantiated superlatives
- Keep all information current and accurate across every channel
A firm that treats compliance as a marketing constraint is missing the point. Compliance is a quality signal, and quality signals are what AI cites.
Read the two lists side by side and the overlap is almost total. Both reward specific, attributed, verifiable statements. Both punish the vague and the unverifiable. That means the work you do for compliance is not overhead sitting alongside your SEO. It is the same work, done once, for two reasons. And it makes the review process easier to run: a page built on verifiable statements needs one check against one rule, not two projects maintaining two versions of the same copy.
The discipline is also easier to hold than it sounds. The test for any sentence on a legal website is whether you could point to the source. A statement about a process can be checked against legislation. A statement about a practice area can be checked against accreditation. A statement about an outcome generally cannot be checked at all, which is exactly why it should not be there.
How 21 Webs Structures Legal SEO
We build law firm websites and run SEO for legal practices across Australia. Here is how our approach works in the AI era.
We start by separating research queries from hiring queries. Before we write anything, we sort a firm’s target keywords by intent. The informational queries get planned as citation assets, because their job is to be quoted rather than visited. The commercial and local queries get planned as competitive pages and profile work, because their job is still to win a local pack. Treating those two sets the same way is the most common mistake we see.
We build practice area pages as citation assets. Each practice area page is structured around the questions prospective clients actually search. The heading asks the question. The first paragraph answers it directly. The rest of the page expands with depth, legislation references and practical guidance. Every page carries LegalService and FAQPage schema.
We build entity authority systematically. We ensure the firm appears consistently across Google Business Profile, Law Society directories, legal directories and business registries. Name, address, phone number, practice areas and credentials are identical everywhere, because that consistency is what wins both the local pack and the citation.
We attribute content to qualified authors. Every legal content page carries the author’s name, qualifications and practising credentials. Attorney schema links the content to a defined entity that AI systems can verify.
We comply with ASCR Rule 36 on every asset. Websites, meta descriptions, Google Ads copy and directory listings are reviewed for compliance. No false specialist claims. No misleading outcome statements. No unverifiable superlatives.
We track citations alongside local visibility. We monitor whether the firm’s content is cited when a client asks an AI system a question, and whether the firm appears in the local pack for the commercial queries that still return one. Those two measurements map onto the two layers the intent split reveals, and they are not the same metric.
The order of that work is deliberate. Entity work comes before content work, because a perfectly structured page attributed to an unclear entity will not be cited. Author attribution comes before expansion, because volume of anonymous pages adds nothing. Intent sorting comes before everything, because until you know which queries are research and which are hiring, you cannot tell which of your pages are supposed to be quoted and which are supposed to be found.
If your firm is starting this work, the sequence we would suggest is unglamorous. Fix the entity first: same name, same address, same details, everywhere. Add named authors to the pages that drive enquiries. Sort your keyword list by intent and be honest about which half is still worth optimising for clicks. Then rebuild the highest-value commercial pages around the questions clients actually ask. None of those steps requires new software or a new website.
Important FAQs
Why do informational legal searches trigger AI Overviews but commercial ones do not?
How do lawyers get cited by AI?
What is legal SEO?
Does my law firm need GEO?
The Bottom Line
We ran 43 Australian Google searches in September 2026, and the pattern was hard to miss. Informational queries returned an AI Overview 19 times out of 19. Commercial or local queries returned one 1 time out of 24.
Within legal, the divide was absolute. Every informational search we tested returned an AI Overview, and not one of the thirteen commercial or local searches did. All thirteen returned a local pack instead.
That is the finding a law firm should act on. The AI layer is eating legal research, not legal hiring. The research queries need to be structured to be cited, because the visit is no longer the prize. The hiring queries still return a map block, and the firm with the complete profile, the steady reviews and the consistent entity data is the one that wins it.
This test is 43 queries, one snapshot, one date, desktop, one location per query. It is a pattern rather than a census, and we have been careful not to claim more than it can carry. But the direction is clear enough to plan around, and the two halves of a legal keyword list now need two different kinds of work.
Is your firm's content ready for AI answers?
We build law firm websites and run legal SEO for Australian practices. If your content is not structured for the way AI answers legal questions, talk to our team while the window is still open.
Sources
[1] 21 Webs, AI Overview presence test: 43 Australian Google searches, desktop, live SERP data (September 2026). One location per query, one run per query.
