This website uses cookies

Read our Privacy policy and Terms of use for more information.

TL;DR

Google published its first official guide to optimizing for AI Search on May 15, and it explicitly retires GEO and AEO as separate disciplines. The guide says AI search optimization is SEO done well. Google I/O 2026 then announced that AI Mode crossed one billion monthly users, with new Gemini 3.5 Flash powering it and Search Agents that monitor topics 24/7 launching this summer. The acronyms got dismantled. The opportunity got bigger. Most law firms still aren't doing the underlying work that Google literally just told them to do.

Game changers

Google published "Optimizing your website for generative AI features on Google Search" on May 15. It's the first official, on-record guidance from Google about what works in AI Overviews and AI Mode. The document lives on Google Search Central and was announced by John Mueller.

The most consequential line. AI search optimization uses the same crawl, the same index, and the same ranking systems as classic Search. There is no separate AI index. The guide explicitly rejects llms.txt files, content chunking, and special AI markup as paths to citation. The work that earns AI citations is the work that earns traditional search visibility, done at a level most firms have skipped.

Then Google I/O 2026 ran May 19 and 20 with numbers that reset the urgency calculation. AI Mode has now surpassed one billion monthly users, with queries doubling every quarter since launch. Gemini 3.5 Flash became the default model in AI Mode globally on the same day. And Search Agents launched as a 24/7 monitoring layer for high-intent topics, available to Pro and Ultra subscribers this summer. That's the demographic that hires law firms for serious matters.

AI search news

AI search startups raised serious capital this month. Exa Labs closed $250 million at a $2.2 billion valuation on May 20, joining Tavily, TinyFish, and Parallel Web Systems in a wave of AI search infrastructure plays. The category is one of the most attractive in consumer AI right now.

OpenAI's self-serve Ads Manager also went live to U.S. advertisers, integrated with every major holding company and ad-tech firm. OpenAI is targeting $2.5 billion in ad revenue this year and $100 billion annually by 2030. Paid placement inside ChatGPT for legal queries is operational and scaling fast.

The bigger picture: AI search platforms are monetizing aggressively, and organic citation visibility is the counterweight that compounds while the paid market is still forming.

The Google guide is forcing a reckoning across the GEO consulting space. Five tactics that vendors have been selling at premium prices got publicly named and dismissed: llms.txt implementations, content chunking, AI-specific markup, and a couple of others.

What replaces those tactics is not glamorous. Server-side rendering, allowing AI crawlers in robots.txt, non-commodity content with original analysis, query fan-out coverage across sub-topics, and E-E-A-T signals that have been the SEO baseline for years. Boring work. Compounding work.

The interesting wrinkle: the AI citation pool extends well beyond the top ten organic results. A page sitting at position 6 or 7 with deeper, more specific content can earn citations that the generic position-1 page misses. For law firms that don't dominate organic rankings, that's the opening.

Source: DemandSphere

Casey’s Take

The Google guide forced an honest conversation I've been wanting to have for a few months.

For two years, the GEO consulting industry has been selling the idea that AI search is a separate discipline requiring separate tools. Special markup. New file formats. Custom optimization frameworks. Google just said, on the record, that none of that is true. AI search is SEO done well. Same index, same crawl, same ranking signals.

There's a temptation to read that as bad news for anyone in this space. I think it's the opposite, and I want to explain why.

The vast majority of law firms are not doing SEO well. They have practice pages written from template by agencies that don't know the practice area. They have blocked AI crawlers in their robots.txt without realizing it. They have generic content that any agency could produce on any firm's site. They have no jurisdiction-specific data, no attorney-attributed authority signals, no recency markers. Google didn't just say "your AI search strategy is the same as your SEO strategy." Google said the bar is the underlying work. And the underlying work is wide open.

I spent ten years in environments where the difference between people who shipped and people who didn't came down to whether they were willing to do unglamorous, repetitive, structural work. The Army teaches you that quickly. The work is rarely the work you wanted to do. It's almost always the work nobody else will commit to. The people who do it consistently win positions the loud operators never reach.

The same logic applies here. The firm that retrofits its twelve most important practice pages this quarter with the structural changes Google literally just outlined will outearn the firm that buys another round of directory placements. Not because of any acronym. Because most firms won't do the work.

Wild cards

Information agents launching this summer are the surface nobody is pricing yet. A high-net-worth prospective client sets up a monitoring agent on a legal topic relevant to a future matter. The agent watches for months. The firm that publishes consistently makes the monitoring set. The firm that publishes sporadically doesn't exist to the agent. This is a publishing cadence argument, not a content volume argument.

Google killing FAQ rich results in SERPs is a smaller story most people are misreading. The dropdown FAQ boxes in search results are gone. FAQ schema as a structured-data signal for AI extraction is still useful. Don't strip the schema. Just don't expect the SERP feature.

Tip you can use today

Read Google's actual guide, because most marketers in your space haven't. The URL is developers.google.com/search/docs/fundamentals/ai-optimization-guide. About 30 minutes to read carefully.

Then pick one practice page on your firm's site. Run it through the five criteria the guide lays out. Is it indexed? Is the content non-commodity? Does it cover the sub-questions a prospective client would actually ask? Does it have proper E-E-A-T signals? Is the content extractable by AI even if the page sits at position 6 or 7 organically?

Document the gaps. That is the work for the next two weeks. Boring, structural, compounding. The firms that do it now will earn citations long after the GEO acronym is forgotten.

If this helped, stay subscribed.

If you want me looking at your firm's AI visibility directly, grab an AI visibility audit at the button below.

Keep Reading