Case study | SEO compounding into AI Mode visibility
How 18 months of law-firm SEO compounded into being the top-cited legal source in Google's AI Mode
The same content library built during 18 months of an SEO retainer is now the single most-cited source in Google's AI Mode answers for the client's target legal queries. It appears in 298 AI-generated answers, more than any government portal, YouTube channel, or international legal reference in the sampled query pool. Classic SEO did not lose value when AI Mode arrived; it doubled as the foundation the client was already sitting on.
- Google AI Mode citations 298 answers #1 source in the sampled pool
- Share of sampled AI answers citing the site ~50% 298 of 599 answers
- Mentions vs government portal 2.5x gov.za 298 vs gov.za's 117
- SEO foundation before compound 18 months of retainer cluster-building
The client
A multi-disciplinary law firm running an active SEO retainer for several years. The practice covers property, family, labour, data protection and civil litigation, and publishes a content library of practical, plain-English answers to the questions clients most often ask. Every article on the site is written to a specific query intent, not as a general topic overview.
The problem
When we started the retainer, the problem was a familiar one: a law-firm site with strong technical legal content but low visibility on the practical questions actual clients search. Eighteen months of SEO work solved that. Then Google's AI Mode arrived and started answering legal queries directly, drawing its citations from a short list of trusted sources. The new question was whether the SEO investment we had already made would compound into that citation set or get flanked by content built specifically for AI extraction. This case study is about how the answer landed.
What made the SEO work compound into AI visibility
None of it was accidental. The features that make classic SEO work in 2026 are, in large part, the features that make content extractable by AI answer surfaces. These were the workstreams that mattered most.
-
Answer-first article structure
Every article opens with a direct legal answer in the first fifty words, followed by depth. Classic SEO best practice, and precisely what LLMs need to lift a short factual answer verbatim into an AI-generated summary.
-
Statute and case-law citations on every article
Every article cites the specific statute or reported case that governs the question. That is authority for search engines and authority for AI models: both weight source-referenced content higher than paraphrase, and legal queries in particular reward it.
-
Author bios with credentials
Every article carries the name of the admitted attorney who authored or reviewed it. This is E-E-A-T for classic SEO and it is the same trust signal AI models use when they decide whether a source is citable on a legal question. Anonymous legal content does not get pulled into AI answers.
-
Structured data markup on every page
Article schema and LegalService schema on every page. These are the semantic labels that both crawlers and LLM training pipelines pick up on, and they cost almost nothing to add at authoring time.
-
Topical clustering as a permanent workstream
Full URL clusters covering entire matter areas rather than one-off articles, plus an annual refresh cadence to keep every piece against current law. Topical authority signals track back to how comprehensively a domain covers a subject; stale legal content burns credibility with humans and LLMs alike.
The compounding effect
The chart below shows the 18-month climb in monthly organic clicks during the first phase of the retainer, from around 2,000 clicks a month to over 10,000. That was the foundation. What the takeaway explains is what happened when Google AI Mode arrived on top of that foundation.
Source: Google Search Console
18 months of SEO work built a legal content library that ranks classic organic AND gets pulled into Google's AI Mode as a primary source. On a fresh sample of AI-generated answers for the client's target queries, the site appears as a source in 298 answers with a combined 147,000 monthly AI search volume. That is more than any other domain in the sample: ahead of YouTube (201 mentions), gov.za (117), Investopedia (61) and sars.gov.za (60), and roughly 2.5 times the mention count of the equivalent government portal.
What that means in practice: for the queries the retainer was originally built to rank on, the client is now the single most-cited source when Google chooses to answer those queries with an AI summary instead of ten blue links. That shift matters most when you look at what it did to classic click volume. AI Mode is intercepting a growing share of the queries this content used to earn clicks on. But the value did not evaporate. It changed shape. The SEO investment paid a click dividend for two years, and it is now paying a citation dividend, from the same URLs, in an answer surface that most competitors are not even instrumented for.
For sites that invested early in extractable content structure, statute references, credentialled authorship and topical clustering, AI Mode compounds the SEO work already done. It does not replace it.
Is your content library ready for AI Mode?
The features that make SEO content extractable by AI models are largely the features of high-quality technical SEO: direct answers, source citations, credentialled authorship, schema markup, topical depth. Book a free audit of your existing content library to see where the extractable-content gaps are, before AI Mode starts making them expensive.
Book a free audit