Google Business Profile for AI Search: Why GBP Is ChatGPT's #2 Citation Source — and the Checklist Therapists Need
In our July 2026 probe of 96 AI answers, Google Maps links accounted for 27% of ChatGPT's local citations. Here's what that means for therapists — and the AI-specific GBP checklist that turns your profile into a citation engine.
The 60-Second Answer
In our July 2026 probe of 96 live AI answers to local therapy queries, Google Business Profile and Google Maps links accounted for 27% of ChatGPT's raw citations — the second-largest citation category behind individual practice websites. GBP is also the complete foundation of Gemini's local grounding, the same engine that powers Google's AI Overviews. And AI use for finding local businesses jumped from 6% to 45% in a single year (BrightLocal, 2026).
Most therapists treat their GBP as a set-it-and-forget-it asset, last touched when they claimed it years ago. That was fine when GBP only influenced Google's 3-pack. It's a different calculation when GBP is driving AI citations on ChatGPT, Gemini, and — indirectly — Perplexity.
This post covers how each major AI engine uses your GBP, the new review floor for AI visibility, and a checklist of the GBP optimizations that actually move the needle for AI search — not the general setup steps, but the AI-specific ones that most practices are missing.
Why GBP Matters Differently for AI Than for Google Search
Traditional local SEO advice said: claim your GBP, get verified, fill in the basics, collect reviews. That advice was correct for Google's 3-pack. For AI search, the bar has moved in a specific direction.
In Google's 3-pack, GBP gets you discovered. In AI search, GBP gets you cited. When someone asks Gemini "find me an anxiety therapist in [city]," Gemini doesn't discover you — it reads your GBP as its authoritative data source, pulls your practice name, categories, services, hours, and review profile, and decides whether you're the right answer to this specific query.
The optimization goal shifts accordingly. Traditional GBP optimization maximized for clicks on map results. AI-specific GBP optimization maximizes for your profile appearing complete, accurate, and specific enough that an AI engine confidently names you as the answer.
The scale of the shift makes this worth prioritizing. BrightLocal's 2026 Local Consumer Review Survey (1,002 US adults, published February 2026) found that AI became the #3 discovery channel for local businesses in a single year — ahead of Yelp and Tripadvisor, behind only Google and Facebook. ChatGPT specifically was used by 31% of consumers for local business recommendations in the past year; Google AI Mode by 23%. Google's own share of local-business discovery fell from 83% to 71%. That's not consumers leaving Google — it's consumers adding AI queries on top of Google, and those AI queries lean heavily on GBP.
Engine by Engine: How Each AI Platform Uses Your GBP
The three major AI engines use GBP in meaningfully different ways. Understanding which engine reads what — and how accurately — tells you where to focus.
ChatGPT: GBP via Google Maps Links at 27% of Citations
In our July 2026 probe of 96 live therapy query answers, ChatGPT included Google Maps or GBP links in 27% of its raw citations — the second-largest citation category, behind individual practice websites at roughly 64%. ChatGPT doesn't have direct API access to GBP the way Gemini does; it accesses GBP data through two paths: the Bing index (which crawls and indexes Google Maps data) and the Foursquare Places API (which syncs with GBP to maintain listing accuracy).
The accuracy issue surfaces here. SOCi's 2026 Local Visibility Index — an analysis of nearly 350,000 locations across 2,751 brands — found that ChatGPT gets business profiles right only 68% of the time. The other 32%: stale addresses, wrong phone numbers, outdated hours. When an AI engine cites your practice but gives a client the wrong phone number, the citation is worse than no citation.
Gemini: Built Directly on GBP — 100% Profile Accuracy
Gemini is the one major AI engine with a direct pipeline into GBP, which is why SOCi found Gemini achieves 100% profile accuracy across their 350,000-location study. Gemini isn't relying on cached web data — it reads the source. This accuracy advantage translates directly into recommendation rates: Gemini recommends 11% of local businesses, versus only 1.2% for ChatGPT — nearly a 10-to-1 difference, which traces directly to that accuracy gap (SOCi, 2026).
For therapists, Gemini is also the public proxy for Google AI Overviews — the AI summaries that appear above organic results on Google Search. Same GBP-grounding logic applies. A well-optimized GBP produces strong Gemini visibility, which produces strong AI Overviews visibility. One asset, three AI surfaces: Gemini, AI Overviews, and the traditional 3-pack.
Perplexity: GBP Matters Less Directly, But Reviews Flow Through It
Perplexity doesn't have a GBP integration. For local therapy queries, it primarily cites practice websites and directories — especially Psychology Today, which dominated our probe data alongside individual practice pages. GBP still matters indirectly, though, through reviews: research on AI local search in 2026 puts Perplexity as using reviews in 100% of its responses to local service queries, and those reviews are often sourced from or cross-referenced against Google's review ecosystem.
A complete GBP with an active, recent review profile raises your trust signals across the whole AI stack — even on platforms that don't directly read your profile data.
The New Review Floor for AI Visibility
The review requirements for AI visibility have shifted in 2026. Pre-AI, therapists could rank in Google's 3-pack for long-tail queries with 5–10 reviews. AI engines apply a higher filter.
Research on AI local search in 2026 puts the competitive minimum at around 30 reviews at 4.3 stars or higher to appear consistently in AI recommendations for competitive therapy categories. This isn't a hard cutoff — no AI engine publishes its algorithm — but it tracks with our own probe data: the practices that appeared in AI answers tended to have review counts in the 20s and 30s, not the single digits.
BrightLocal's 2026 survey adds the consumer lens: 47% of consumers won't use a business with fewer than 20 reviews, and 31% only use businesses rated 4.5 stars or higher. AI engines optimize for those consumer preferences — they surface the businesses consumers would be most likely to trust.
If your GBP has fewer than 15 reviews, closing that gap is your highest-ROI AI discoverability action — ahead of optimizing categories, adding services, or uploading photos. Review volume is the entry ticket. See our guide to getting Google reviews as a therapist for the specific ask sequence that works in a therapy context.
One additional nuance: AI engines weight review recency, not just volume. A practice with 40 reviews all from 2022 typically scores lower in AI recommendations than one with 20 reviews including 5 from the past 60 days. The goal isn't a one-time push to a number — it's a steady trickle. Even one new review per month is better than a burst-and-silence pattern.
The AI-Specific GBP Checklist
The following checklist goes beyond standard GBP setup (claim, verify, fill in NAP). It targets specifically how AI engines read GBP — the fields they extract, the signals they weight, and the content formats that produce citations.
| GBP Element | AI Relevance | What to Do |
|---|---|---|
| Primary category | Highest — AI uses this to match specialty queries | Use the most specific option: "Marriage & Family Therapist," "Psychologist," "Mental Health Clinic" — not "Healthcare Provider" |
| Secondary categories | High — each adds a matching signal | Add all relevant categories: "Child Psychologist," "Counselor," "Therapist." Gemini reads all of them. |
| Services section | High — Gemini reads service names to match modality queries | Add every modality and specialty as a separate service with a one-sentence description: "EMDR Therapy," "Couples Counseling," "Anxiety Treatment" |
| Business description | Medium-high — AI summarizes this text | 750 characters: lead with what you do and who you serve, include the city, name your top 3 specialties, end with credentials. No fluff. |
| Q&A section | Medium — AI can cite Q&A answers verbatim | Seed 5–6 questions clients actually ask; answer them yourself. "Do you offer telehealth?" "What insurance do you accept?" "What modalities do you use?" |
| Review volume + recency | Highest — used in 100% of Perplexity and 58% of ChatGPT local responses | Target 20+ reviews, 4.3+ average; at least 1 new review per month |
| Review responses | Medium — signals active, engaged practice | Respond to every review within 7 days, positive and negative |
| Photos | Low-medium — proxy for active business | Minimum 10 photos: exterior, waiting room, headshot, credentials; update at least 2 per year |
| Monthly posts | Medium — indexed by Gemini and Google | One post per month: 150 words, question headline, direct answer in first sentence. "What is EMDR therapy?" style. |
The business description gets overlooked most often. Most therapist GBP descriptions read like: "I'm a licensed therapist helping individuals and families work through life's challenges." That's 12 words of actual information. AI engines have nothing to work with. Rewrite it: "Marriage and family therapist in [City] specializing in anxiety, trauma, and couples counseling. EMDR-trained LMFT with 10 years of experience, accepting new clients and most PPO insurance plans." Now Gemini can match you to specific queries.
The Accuracy Gap: Why ChatGPT Gets Your GBP Data Wrong 32% of the Time
SOCi's finding that ChatGPT is accurate on business profiles only 68% of the time explains something therapists encounter regularly: ChatGPT sometimes cites them with an old address, an outdated phone number, or wrong hours — information the therapist corrected on GBP months ago.
This happens because ChatGPT doesn't read GBP directly. It reads the cached version of your GBP data in the Bing index (which may be weeks or months behind), plus any web pages that mention your practice: old directory listings, a website page not yet updated, a press mention from three years ago. When those sources conflict with your current GBP, ChatGPT guesses which is correct — and it guesses wrong nearly a third of the time.
The fix requires auditing every source, not just updating GBP:
- Search Bing for your practice name. Bing's index is what ChatGPT actually reads. Note any outdated information in the results.
- Check your Bing Places listing directly at bingplaces.com — it should match your GBP exactly. Discrepancies here propagate into ChatGPT answers.
- Google your practice name and audit every result: old directories, cached addresses, any page where your NAP appears.
- Check your website header, footer, and contact page. Every place your address or phone appears should match GBP exactly — including suite numbers and formatting.
AI engines cite the most consistent version of your information. A therapist with perfect NAP consistency across 8 platforms will be recommended more often than one with a perfect GBP and conflicting data everywhere else. Our anatomy of AI citation analysis covers how each source type contributes to that consistency score in detail.
How Often to Update Your GBP for AI Visibility
AI engines weight content freshness. Review recency, post recency, and even the dateModified signal in your website's schema all affect citation confidence. Gemini reads your GBP directly and sees recency in real time. ChatGPT reads Bing's cached version — which typically lags 30–60 days. That lag has an implication: changes you make today won't reach ChatGPT recommendations for over a month. Update early, update consistently.
A practical maintenance cadence:
- Monthly: One GBP post, 150–200 words, one direct question answered. Respond to any new reviews within 7 days.
- Quarterly: Review your Services section and business description. If you've added a modality, a new location, or changed insurance networks, update immediately — and update your website too so the sources stay consistent.
- Annually: Audit your photos (add 2 new ones), re-read your Q&A section for anything stale, and run our AI visibility audit prompts to test what Gemini and ChatGPT are actually saying about you.
The 30–60 day Bing lag also means GBP is not where you respond to an AI-accuracy emergency. If ChatGPT is citing wrong information about your practice, the fastest fix is to update Bing Places first (direct impact on ChatGPT), then your website, then GBP. GBP is the authoritative source, but Bing is ChatGPT's actual data feed.
What Your GBP Can't Do Alone
GBP is the second-largest AI citation source for local therapy queries — but it isn't the first. Individual practice websites accounted for roughly 64% of all classifiable AI citations in our probe, versus the 27% contributed by GBP and Google Maps links. A strong GBP without a strong website will consistently lose AI citations to practices that have both.
The relationship between the two is reinforcing, not competing. Your GBP categories and service names should mirror your website's service page slugs and headings. Your business description should use the same phrasing as your website's homepage and "About" intro. When AI engines see matching language across your GBP and your website, they treat you as a more authoritative, consistent entity — and cite you more.
The checklist in this post covers the GBP side. For the website side — the service page structure, schema markup, and content formatting that drive that 64% — see our complete GBP setup guide for the traditional SEO layer, and the anatomy of AI citations post for the full source breakdown.
At WebsiteTherapy, every client site's practice identity — name, address, phone, specialties, hours — is stored once in a central data file and syncs automatically to GBP, Bing Places, Apple Maps, and Foursquare. When a therapist tells our AI assistant "update my hours," it updates their website and all four platform listings simultaneously. We also run weekly AI probe monitoring so clients can see whether ChatGPT and Gemini are surfacing them for their key queries — and which GBP signals are holding them back when they're not. See how it works or browse the features.