Google AI Mode vs AI Overviews vs ChatGPT: How Each Picks a Lawyer (2026)

Short answer: All three read the open web, but they choose sources differently. Google AI Overviews summarise a normal Google result set and lean on pages already ranking plus Google Business Profile data. AI Mode fans a question out into several searches and rewards pages that answer sub-questions in depth. ChatGPT runs its own web search, cites far fewer sources per answer, and weights directory profiles, reviews and mentions heavily. A law firm that fixes entity signals, publishes answer-first pages and builds reviews shows up in all three; a firm that only chases Google rankings shows up in one.

Why does the distinction matter for a law firm?

Because the same client may use all three in one afternoon. In August 2026, 41.9% of surveyed consumers said they would use ChatGPT to research which lawyer to hire; Google use fell to 71.9% in the same survey (iLawyer Marketing). AI Overviews now appear on roughly a quarter of legal queries on Google (SEO Discovery, 2026), and AI Mode is expanding from opt-in to default for many searches. Each surface names three or four firms and stops. Being one of them in one surface and absent from the other two is the most common pattern we see in Los Angeles audits.

How does Google AI Overviews pick sources?

AI Overviews sit at the top of a normal results page and are generated from that page's candidate set plus Google's knowledge graph. Three things follow:

• Ranking still matters, but less than people think. 29.8% of domains cited in AI Overviews are not on page one for the query (Digital Applied), and only 38% of AI Overview citations come from the organic top ten (Soar). The overview reaches down the results for the page that answers the question best.
• Local queries pull from Google Business Profile. For "family lawyer near me in Burbank", the overview is built alongside the map pack, so the profile's category, reviews and consistency carry into the AI answer.
• Answer position wins. About 55% of cited passages come from the first 30% of a page. A practice page that opens with two paragraphs of firm history loses to one that opens with the answer.

What to build: a Google Business Profile with the right primary category and steady reviews, and practice pages whose first paragraph answers the question the client typed.

How does Google AI Mode pick sources?

AI Mode is a separate conversational surface. It takes the question, breaks it into several sub-queries ("query fan-out"), runs them, and writes a longer answer with a source panel. Compared with AI Overviews:

• Depth beats position. A page that covers custody, support, timelines and costs in Los Angeles County can be cited for a sub-query it never targeted, because the fan-out found the matching section.
• Follow-ups keep you in or out. Users refine ("what about at the Van Nuys courthouse?"). Pages that answer the follow-up win the follow-up. This is why courthouse pages earn citations quickly.
• Entities are checked across sources. AI Mode is more likely than AI Overviews to cross-reference a firm's name with directory profiles before naming it.

What to build: intent-complete hubs, one per practice area, with question-phrased H2s and a section per sub-question, plus the city and courthouse pages the follow-ups reach for. See the GEO playbook for Los Angeles.

How does ChatGPT pick sources?

ChatGPT with search runs its own web queries through OAI-SearchBot's index and fetches pages on demand. Its behaviour differs from Google's in three measurable ways:

• It cites rarely. One citation index found brands cited in about 0.6% of ChatGPT answers versus around 13% for Perplexity (5W). Each ChatGPT citation is scarce, which makes it valuable when it lands.
• It concentrates on a few domains. Roughly 68% of citations across engines flow through about 14 domains (5W citation index; Leapd): directories, Reddit, Wikipedia, YouTube, LinkedIn and major publishers. For a "which lawyer" question that means Avvo, Justia, FindLaw, Yelp and the State Bar profile carry as much weight as the firm's site.
• Mentions beat links. Ahrefs' 75,000-brand study: unlinked web mentions correlate with AI visibility at 0.66, YouTube mentions at 0.74, backlinks at 0.22.

What to build: directory profiles that match the website exactly, a review programme on Google and Avvo, named attorney profiles on LinkedIn, and the entity signals that let the model confirm the firm exists.

Side by side

What works in all three?

The overlap is the program, and it is the same one we describe on the AI search visibility pillar:

1. Consistent name, address, phone and practice list across the website, the State Bar, Avvo, Martindale, Justia, Yelp, Google and LinkedIn.
2. LegalService and Person schema that link those profiles with sameAs.
3. One answer-first page per practice area, per city and per courthouse the firm appears in.
4. Reviews requested at every matter milestone, within bar rules.
5. Pages in the language the client asks in; AI Mode and ChatGPT both answer Spanish, Armenian and Korean questions from sources in those languages.

What does not transfer?

• Backlink campaigns. Links correlate weakly with AI visibility (0.22) and not at all with ChatGPT citations of small firms.
• Keyword-stuffed city pages. AI Mode's fan-out looks for answers to sub-questions, not for the phrase "best divorce lawyer Burbank" repeated eleven times.
• Paid placement. None of the three sells organic citations. ChatGPT ads exist and are labelled; they are a separate budget line and not what this article is about.

How to see where your firm stands in each engine

Run twenty of your clients' real questions through each surface, in the language they ask them, and record which firms are named. We do this for 40 to 60 questions across ChatGPT, Perplexity, Gemini and Google AI Overviews and AI Mode, repeated to control for randomness, in the free AI visibility check. The report tells you which surface you are missing from and why the cited firms win there.