A four-client pilot by Denver-based Behavioral Health Partners recorded an average increase of more than 60% in AI mentions after one month of work per client, the agency announced on September 30, 2026, alongside the launch of its Generative Engine Optimization (GEO) service for addiction treatment centers.

The figure comes from the company's own internal before-and-after testing — not from an independent study — and it measures how often client centers appear in AI-generated answers. It does not represent growth in admissions or patient outcomes, a caveat the agency itself states plainly.

What did the testing involve?

Behavioral Health Partners began testing its GEO approach in May 2026. The company observed its first substantial movement in July and continued recording gains through August and September before going public with the service.

The tests compared responses to identical prompts before and after implementation, run across 12 configurations spanning four AI platforms: ChatGPT, Gemini, Claude and Perplexity. Configurations covered different access levels and response modes, including unsigned-in access where available and reasoning-focused modes.

According to the company, repeated testing showed participating centers appearing more frequently in responses relevant to their states, the services they provide and the credentials they hold. Client identities remain confidential.

Why target AI search for treatment queries?

The service addresses a shift in how families search for addiction treatment. Founder and Managing Partner Adam Vibe Gunton described the difference in behavioral terms.

"A parent might type 'drug rehab near me' into Google. With AI, they may explain what they're seeing in their son, describe the behaviors that worry them, and ask how to help," Gunton said. "When treatment is an appropriate next step, the information behind that recommendation needs to accurately reflect what a center does and who it can help."

The goal is accuracy as much as visibility: ensuring AI answers about a center's locations, services, clinical teams and credentials match reality.

How does the seven-step process work?

The service combines on-site changes with external source work:

• Direct-answer content — restructuring key pages to answer questions about treatment, programs, insurance and the admissions experience
• Structured data — schema markup describing the organization, locations, services and people, plus technical improvements supporting content access
• Credential visibility — presenting staff roles, licenses, accreditations and clinical review clearly
• Listing reconciliation — fixing inconsistencies across websites, business profiles and directories
• Third-party citations — building accurate references through publications, directories and professional organizations
• Proprietary off-page methods — citations, articles and website partnerships focused on relevant treatment questions and geographic markets
• Monthly measurement — repeating prompts, tracking mentions and citations, and reviewing how AI responses describe each center

The offering extends the agency's existing SEO work and adds AI answer visibility and accuracy to the metrics used to evaluate a treatment center's online presence.

What should providers do now?

Gunton, whose background includes addiction recovery and treatment-center development, framed the stakes around message accuracy rather than raw rankings.

"Our goal is for someone who needs help to find the right treatment center, a place where they feel safe, welcome, and like they were meant to find it," he said. "We've worked to connect families with that message through Google. Now we're bringing that same purpose to AI search."

He urged treatment providers to begin evaluating their presence in AI answers immediately, including whether those answers accurately describe their programs and qualifications. "This is still early, and treatment centers need to take it seriously," Gunton said. "The work starts with making sure these systems can find accurate, consistent information about who you are and the people you serve."

For the behavioral health vertical, the launch signals that GEO is moving from enterprise experimentation into regulated, high-stakes categories where an inaccurate AI answer carries real consequences. The 60% figure is a vendor-reported number from a tiny sample of four clients, so treat it as directional rather than conclusive. Providers and agencies in this space should monitor how consistently ChatGPT, Gemini, Claude and Perplexity describe their programs and credentials as these systems continue to reshape how families find care.