Automation, Attribution & AI: The Tech-Driven Path to Growth in Behavioral Health

Steve Donai: There are 7,500 treatment centers in America, give or take. A good 7,000 of them probably don't have the foundations in place for a basic tech stack. Gary, let's kick it off — what would you start any treatment center with?

The foundational tech stack every treatment center needs first03:00

Gary Garth: Try to get the fundamentals in place. Everybody wants the newest Tesla, but are we tracking everything? Do you have a CRM? A lead rating system? Are you following up on opportunities effectively? Start with a call tracking provider — call analytics integrated into your marketing channels so you have attribution and can compare apples to apples. That's the foundation. And don't worry about AI and all these sexy tools, because if you don't have SOPs and processes in place, they're absolutely useless. Get the one-on-one in place, understand your attribution across every channel. Then you can optimize from there.

Keegan Warrington: Even just your website — do you have Google Analytics connected? Google Search Console? Setting up foundational tracking tools to understand website behavior is step one before any of the advanced stuff.

Attribution, dashboards, and which KPIs actually matter07:00

Steve Donai: Industry average right now is 30 to 40% of all calls are missed. If you want a quick 30 to 40% ROI — just answer the phone. That's it. A CRM, call tracking metrics, making sure the phone's being answered, when, and who's answering it. Wave a magic wand, that's what I'd fix first.

Gary Garth: The KPI that matters most? Cost per viable VOB. Hands down. When we did that test with Sandstone — optimizing the landing page and passing that signal back to Google — we were telling them: find us more of these kinds of prospective patients. Google knows a ton about your audience that we don't. Making sure you're collecting that data and feeding it back is extremely valuable. And when you can compare Google Ads, SEO, Recovery.com, Microsoft Ads, and BD using the same four KPIs — cost per lead, cost per VOB, cost per viable VOB, cost per admission — now you can make real decisions about where to put your money.

Keegan Warrington: I have a story about this. A detox center was spending $25,000 a month on Google Ads. No CRM, no lead tracking. We set up HubSpot, set up proper attribution, and found out the Google spend wasn't driving viable leads at all. We turned it off. Admissions didn't drop. The client saved several hundred thousand dollars a year — and could reallocate that money to channels that actually worked.

Using data to reallocate budget and make better marketing calls12:00

Steve Donai: We've had similar ones. You turn off the paid spend, the admissions team is terrified — and then nothing happens. They're like, "finally, we can talk to someone for more than five seconds."

Gary Garth: On automation — the most valuable one I'd put in every treatment center is a text drip to closed-lost and MIA leads. Short, human-feeling messages. "Just checking in, how are you?" You're not gonna get responses for 17 messages. But all of a sudden when they're ready, a few reply. In Q4 when it's challenging, that automation is the difference. You can't do retargeting in this industry — this is your alternative.

Keegan Warrington: My magic wand automation? Asking for reviews. Easy to set up, three high-value impacts: conversion rate improvement, SEO lift, and marketing assets you can use everywhere. Most facilities are not doing this consistently.

Aligning leadership, marketing, and admissions around shared metrics18:00

Gary Garth: We take it even further — we have a physical planner, the Gold's Grid & Greatness Planner, that we white-label for facilities. You hand it to the patient at discharge during the "honeymoon phase." There's a QR code to leave a Google review. It feels like a gift — reciprocity creates the obligation to review. Set a KPI of 20–25% of discharges leaving a review. People will say you're crazy. But if leadership owns that KPI, it happens.

Steve Donai: On AI — are you using it, and where are you seeing traction?

Automation strategies for admissions follow-up22:00

Keegan Warrington: I'm using AI in basically every aspect of the business — content creation, dicing up videos, turning blogs into audio with ElevenLabs. Outside of handling medical information directly, I'm open to trying AI on almost everything.

Gary Garth: Not for content creation — that's a huge no. Google wants authenticity, expertise, authority, trustworthiness. But what we do is create the "voice of elev8" — feed AI our best-performing emails, text messages, follow-up sequences, everything that's actually worked — and then use it to generate follow-up collateral that's templated but genuinely ours. Not ChatGPT writing content. You feeding it your information and leveraging it at scale.

Gary Garth: On AI admissions reps — I'm working with two or three vendors building them, and I'm not open to that. There's too much at stake. AI can give the wrong advice and that's a person's life. Part of me says 40% missed calls and you just plug in the agent. But I'm not convinced there's no risk. Not yet.

Steve Donai: Same. I don't wanna be the agency whose client has a patient call in at a critical moment — a real crisis — and the tool wasn't ready for that severity. I've been on those calls over the years. It's hard to imagine a computer getting it right now. I think we'll get there. I just haven't seen it yet.