Meta AI for Therapists: How to Get Found on 1.2 Billion Devices Across Facebook, Instagram & WhatsApp

The number deserves a pause. Meta AI — the AI assistant embedded across Facebook, Instagram, WhatsApp, Messenger, and the standalone meta.ai platform — reached 1.2 billion monthly active users in 2026. That makes it the world's second-largest AI assistant behind ChatGPT, and it got there in 18 months: the fastest growth trajectory of any AI platform in history (DemandSage, 2026).

For therapists mapping where potential clients are now searching for mental health providers, that scale is significant. Especially when you factor in the distribution: 63% of Meta AI engagements happen on WhatsApp — the messaging app where hundreds of millions of people already ask questions and get answers in a conversation thread alongside their actual friends and family (DemandSage, 2026).

Then on June 15, 2026, Meta pushed the stakes further. The company launched Facebook AI Mode globally — a new search experience embedded directly in the Facebook app that answers local recommendation questions by synthesizing what people say publicly across Facebook Groups, Reels, and public posts. Not a list of links. Not a directory. An AI-generated answer drawn from real conversations already happening on the platform, powered by Muse Spark, the model Meta Superintelligence Labs introduced in April 2026.

For therapists trying to be found, this represents a meaningful new channel with optimization mechanics that are fundamentally different from Google AI Mode, Perplexity, or Apple's Siri AI — and a user base that overlaps heavily with the demographics most likely to seek mental health support. Most practices have no presence there at all. That gap is addressable, and this guide covers exactly how.

Meta's AI infrastructure isn't a single product. It's an overlapping ecosystem of surfaces, each with different data sources and different optimization levers. Understanding which surface works which way prevents wasted effort.

The through-line across all five surfaces: public, structured, complete information about your practice. What's different from other AI platforms is that "public information" on Meta includes social content — posts, community discussions, Reels — not just directories and websites. That creates optimization work that Google, Apple, and Microsoft don't require.

Of the five surfaces above, Facebook AI Mode is the most structurally significant for local service providers. It launched globally on June 15, 2026, powered by Muse Spark — a model from Meta Superintelligence Labs that Meta positioned as its most capable to date. The feature is accessible through the existing Facebook search bar, appearing as a new tab that routes recommendation and opinion queries to AI-generated answers instead of link lists.

The defining characteristic of AI Mode is what it draws from. Unlike Google AI Mode (which draws from Google's web index and Knowledge Graph) or Copilot Health (which queries provider directories), Facebook AI Mode synthesizes answers from content that exists within Meta's own social ecosystem: public Facebook posts, Group discussions, Reels clips, and public community conversations. When someone types "trauma therapist near me" or "who does anyone recommend for anxiety therapy in my area?" into Facebook search, AI Mode generates an answer from what real people have been saying publicly on the platform.

This creates a dynamic with no equivalent on other AI platforms. A positive mention of your practice in a local mental health support Group post becomes retrievable material for AI Mode recommendations. A public Facebook post where you answer a question about EMDR therapy adds to the corpus AI Mode can draw from when a potential client asks about EMDR therapists in your city. An inactive Page with no posts from the last year provides almost nothing for AI Mode to work with.

WhatsApp is where the volume lives. Meta AI on WhatsApp accounts for 63% of all Meta AI engagements — more than Facebook, Instagram, and Messenger combined. The format feels natural to users: they're already in a chat interface, and asking Meta AI for a local service recommendation is one small step from texting a friend for the same thing.

When a WhatsApp user asks Meta AI to suggest a therapist in their area, the assistant draws from a combination of its training data and web-indexed information. The Muse Spark model underlying Meta AI has broad general knowledge, and for local business queries it supplements that with public business data — which is why your Facebook Business Page information (which flows into Meta's business data ecosystem) and your website's indexed content both become relevant.

On Instagram, Meta AI features are integrated into the app's search and discovery layer. Public Reels and posts from a business profile contribute to Meta AI's understanding of what that practice offers and who it serves. An Instagram Business profile with active public content — even one cross-posted from Facebook — is a supporting signal in Meta AI's overall knowledge of your practice.

On Messenger, the new Meta Business Agent (launched June 3, 2026) enables practices to run an AI-powered assistant that responds to inbound messages automatically. More than 1 million businesses are already using Meta Business Agent on WhatsApp and Messenger (Meta, 2026). For a therapy practice, this is the conversion layer: a potential client who found you through a Meta AI recommendation and reaches out via Messenger or WhatsApp needs a prompt, professional response. Business Agent provides that without requiring you to be available around the clock.

Meta has been more explicit about Facebook AI Mode's data sources than about the general Meta AI assistant's sources. Here's what's confirmed and what it means for optimization priority:

The hierarchy tells you where to invest. Facebook AI Mode runs almost entirely on social content — not on directories, not on your website. Meta.ai and the WhatsApp assistant also draw from web-indexed content, which means your website matters more for those surfaces. A practice that does both — active public social presence and well-structured website content — covers the full stack.

A 2026 BrightLocal survey found that 45% of consumers have used AI to find local business recommendations, up from just 6% the year prior — a more than 7× increase in a single year (BrightLocal, 2026). The practices getting in front of that shift now are building early-mover advantages that will compound as usage continues to grow.

Every step below is free. The setup work runs about two to three hours total. The ongoing investment is a regular posting rhythm — once or twice a week across Facebook, with monthly participation in one or two Groups.

Your Facebook Business Page is the entity Meta AI recognizes as your practice. It's the foundation for AI Mode recommendations, the Business Agent configuration, and the WhatsApp Business connection. Without a complete Page, you're structurally absent from Meta's AI ecosystem regardless of how active your personal Facebook profile is.

What to complete:

• Business name: Your legal practice name, exactly as it appears on your license and website. Inconsistencies between your Page name, website, and Google Business Profile reduce trust signals across all platforms.
• Category: Choose the most specific applicable option — "Mental Health Clinic," "Psychologist," "Therapist," or "Marriage Counselor" — not generic "Health & Wellness." Meta AI uses category to match your Page against specialty-specific queries the same way Apple Maps and Google do.
• About section: Write 200-400 words covering your specialties, therapeutic modalities, populations you serve, and insurance you accept. This natural-language description is what Meta AI reads when generating practice recommendations beyond the category match.
• Services section: List each modality individually — EMDR, CBT, DBT, somatic therapy, Gottman Method, couples therapy, LGBTQ+ affirming care. Individual service entries affect which specialty sub-queries ("therapist who does EMDR near me") your Page can match.
• Contact information: Phone number, website URL, and appointment booking link — current and matching every other listing you have. Your Google Business Profile, Apple Business Platform, and Facebook Page should show identical NAP (Name, Address, Phone).
• Hours: Current and granular. Include telehealth-only days and whether you're actively accepting new clients. Meta AI personalizes local recommendations in part by availability context.

For telehealth-only practices: mark your Page as a service-area business rather than a physical location. Don't list a home address or virtual office — it creates trust and potential compliance issues, and Meta's systems handle service-area businesses appropriately for local recommendations.

Facebook AI Mode draws from recent public content. A Page with no posts from the last several months is a weak retrieval source — there's little for the model to synthesize from when someone asks about therapists in your specialty area. A Page with consistent public posts covering your clinical areas is actively building an indexed body of material that AI Mode can surface in recommendations.

Two to three public posts per week is sufficient. More important than frequency is topic relevance:

• Educational posts about your specialty areas: What is EMDR and who benefits from it? What's the difference between CBT and DBT? How do you know if couples therapy is right for your relationship? These become the reference content AI Mode synthesizes from when someone asks about therapists in your modality.
• Answers to common intake questions: What does a first therapy session look like? Do you offer sliding scale fees? How long does a typical therapy engagement last? Posts that answer these directly create retrievable Q&A material.
• Practice updates: Currently accepting new clients. New insurance accepted. Evening slots opened up. AI Mode indexes availability-relevant signals.
• Awareness month content: July is Minority Mental Health Month. Posts recognizing awareness months relevant to your practice areas add contextually relevant content to your public corpus during the weeks when related searches spike.

Format matters too. AI Mode can pull from text posts, Reels, and Marketplace listings. Publishing occasional Reels — even brief 60-second videos addressing a common therapy question — gives Meta AI additional retrieval formats beyond text, which affects how your practice is represented in different query types.

This is the highest-leverage optimization step unique to Facebook AI Mode — and the one most therapists skip entirely.

Facebook Groups are explicitly one of the primary corpus sources that AI Mode synthesizes from. When community members ask "can anyone recommend a good anxiety therapist in Denver?" inside a local mental health support Group, or "does anyone have experience with trauma-informed therapists in this area?" inside a parenting community, those posts and the responses they receive become retrieval material. When AI Mode generates a recommendation for a user asking a similar query, it can surface content from those public Group discussions.

Genuine, helpful participation in public Groups where your potential clients gather — not promotional posting, but actual clinical expertise offered freely — builds the kind of community-endorsed presence that AI Mode draws from. This looks like:

• Identifying 3-5 public Facebook Groups relevant to your specialty (local mental health forums, parenting groups for populations you serve, community wellness groups)
• Answering clinical questions with real, substantive expertise — not "DM me for more information," but actual useful responses
• Introducing your professional background when relevant without making it promotional
• Maintaining consistent monthly presence rather than occasional bursts

The ethical line is clear and worth being explicit about: share expertise, don't solicit clients in Groups. Professional conduct in public online spaces is subject to the same ethical standards as any other public professional context. But genuine participation that happens to demonstrate your expertise publicly is exactly what Facebook AI Mode is designed to surface.

Instagram public content feeds Meta AI's general assistant and the AI features Meta is progressively rolling into Instagram's search experience. A complete Instagram Business profile with active public content — posts, Reels, a bio that describes your specialty and populations — is a supporting signal in Meta AI's overall picture of your practice.

For practices already posting to Facebook, cross-posting to Instagram adds minimal overhead while expanding your presence in Meta's content ecosystem. The bio is worth particular attention: 150 characters that include your specialty, credential, and location give Meta AI's crawl a clean, structured signal. "Licensed LCSW specializing in trauma & anxiety | Downtown Portland | DM to book" is more extractable than "Welcome to my practice page."

Instagram is a supporting signal, not the primary lever. The Facebook Page and Groups work drives more AI Mode impact. But given that Instagram content explicitly feeds the Meta AI ecosystem, maintaining a public business profile costs little and compounds over time.

WhatsApp is where 63% of Meta AI engagements happen. A potential client asking Meta AI on WhatsApp to recommend a therapist in their area is one of the highest-intent pathways in this entire ecosystem — they're in a private conversation, they're asking for a recommendation, and they're ready to follow through.

Setting up WhatsApp Business (free) gives your practice an official presence in the platform where most Meta AI conversations happen. The WhatsApp Business profile fields — business description, category, hours, website link, booking URL — are the structured data layer that Meta AI's WhatsApp surface draws from for practice-specific recommendations.

Meta Business Agent, launched June 3, 2026, extends this by enabling an AI assistant to respond to inbound WhatsApp and Messenger messages on behalf of your practice. More than 1 million businesses are already using it. For a therapy practice, the configuration is straightforward: set up responses to common initial inquiries (availability, specialties, insurance, how to book) and enable routing to you directly for anything requiring clinical judgment.

The conversion rationale is simple. A potential client who found your practice through a Meta AI recommendation and followed through to send a WhatsApp message is showing high intent. A slow or no response at that moment loses the lead. Business Agent ensures a prompt, professional response is available even outside business hours.

For meta.ai's standalone assistant and for Meta AI on WhatsApp, your website remains a relevant signal. The general Meta AI assistant has web access and draws from indexed content alongside its training data when generating local provider recommendations. The same structured data that serves Google, Apple, and Bing also serves Meta AI's web-facing surfaces.

The highest-impact structured data for Meta AI's web layer:

• LocalBusiness or MedicalBusiness schema: Practice name, address, phone, hours, and specialties in machine-readable format — the structured equivalent of your Business Page data but on your own domain.
• Person schema with hasCredential: Your license type (LPC, LCSW, LMFT, PsyD, PhD), modality certifications, and educational background. This is the trust signal that distinguishes a licensed professional from a wellness blogger in Meta AI's knowledge model.
• FAQPage schema on service pages: Each modality or population you serve, structured as questions and answers. These are the highest-extraction format for AI assistants answering follow-up questions about a practice.

Verify that Meta's crawler isn't accidentally blocked in your robots.txt. This is a common issue on older therapy websites with restrictive default CMS configurations, and it silently removes your site from the web-crawl layer across every AI platform simultaneously. For a full walkthrough of AI crawler configuration, see our guide to robots.txt for therapist websites.

Meta AI is the sixth major AI discovery surface therapists now need to consider. The complete picture, with each platform's primary data source and optimization lever:

The common thread across all six: accurate, complete, publicly accessible information about your practice. The Meta AI-specific layer — active public social presence and Group participation — requires a different kind of ongoing content effort than the directory and structured data work the other five platforms emphasize.

The practical implication: the foundational work (complete GBP, complete Apple Business Platform, updated NPI record, structured data on your website) serves five of the six platforms simultaneously. Meta AI's social layer is additive — it rewards therapists who are willing to be publicly present on the platforms where conversations about mental health providers are already happening.

For the broader context of how all six platforms fit together — and which data sources overlap versus which require separate work — see The Therapist's Complete Guide to AI Discoverability. For the closest equivalent optimization guide in terms of format and investment level, the Google AI Mode guide covers a similar depth of platform-specific work.

At WebsiteTherapy, every practice site ships with the structured data — LocalBusiness, Person, FAQPage — that the web-facing surfaces of Meta AI, Google, Apple, Bing, and Perplexity use to understand and recommend your practice. The social layer on Meta's platforms is something each practice manages directly. But the technical foundation that five of the six AI channels draw from is handled at the platform level, so it doesn't require manual upkeep as your site changes. See how it works or compare what this foundation looks like against the alternative in our Psychology Today comparison.

Sources: DemandSage, "Meta AI Users Statistics (2026) — Weekly & Daily Active Users" (2026); Meta Newsroom, "Be There for Every Customer With Meta Business Agent" (June 3, 2026); TechCrunch, "Meta's new 'AI Mode' on Facebook pulls from public info across its platforms" (June 15, 2026); Digital Applied, "Meta AI Mode in Facebook Search: Brand Visibility 2026" (2026); TechWyse, "Facebook AI Mode Pushes Meta Deeper Into Search" (2026); BrightLocal, Local Consumer Review Survey (2026), as cited by The Big Fish SEO; Meta Newsroom, "Introducing Muse Spark: Meta's Most Powerful Model Yet" (April 8, 2026).