When a conversation turns to AI, people often voice concern about the technology taking over or automatically making critical, even life-and-death, decisions. For behavioral health professionals, the reality of tapping AI-powered tools for treatment centers is remarkable but practical.
AI is an efficiency tool for a department. Instead of replacing people, AI done right often takes on the tasks associated with tactical work that employees have. That, in turn, frees workers to focus on elements of their job with the highest value.
Here’s Where AI Makes a Difference
A behavioral health organization can see immediate gains from AI when deployed for operations, specifically admissions. This is the part of the organization where volume, speed, and follow-through are critical. Missed calls in behavioral health aren’t just marketing statistics; they mean a person in crisis doesn’t get connected to care.
Let’s say a treatment center owner staffed their call center with 50 people. The call center agents would likely have one, maybe two, peak hours where 60 phone calls happen simultaneously. In other words, the owner can’t necessarily predict what call volume will be at any given time. Call centers are overstaffed during certain hours, and during peak hours, staff will miss calls. Missing calls affects not only care but also revenue.
Every missed call can mean up to $50,000 in lost revenue (i.e., average per-episode reimbursement). Every pre-admission assessment that takes 20–45 minutes can cost $50 an hour, but more importantly, it takes up a call center agent’s valuable time that could be spent answering calls from other potential patients — calls that may get missed because they are filling out a long medical and behavioral questionnaire. The traditional (non-AI) solutions used by behavioral health facilities range from answering services to on-call staff. The former is impersonal and lacks clinical screening, while the latter leads to inconsistent quality and burnout among staff. The result of these inadequate solutions is often the same: the patient goes somewhere else.
The behavioral health industry is already at a point where AI voice agents can handle overflow calls, field inquiries, collect basic information, and route callers to staff. AI agents can also automate follow-up on “open opportunities,” or potential patients who inquired about help but ultimately did not enter a treatment center.
Research that our organization has done on call volume suggests that admissions teams do not follow up on 90 percent of open opportunities. An AI agent, by contrast, can make thousands of outreach calls and handle hundreds of simultaneous conversations. Nothing falls through the cracks when a facility owner applies AI in this way. An organization that uses AI can add an unlimited number of assistants to its workforce.
Beyond admissions, AI could play a role in alumni outreach. Think about automating check-in calls with former clients, flagging anyone who wants a live conversation, and keeping tabs on those who might fall off an organization’s radar after discharge due to a lack of outreach.
There’s also a role for AI in clinical documentation. There are AI tools on the market now to help therapists expedite progress notes and administrative tasks, freeing them to spend more time listening to and interacting with clients.
Where AI Falls Short
Despite the wide array of applications for AI, the technology won’t replace the way therapists and clients relate to one another. Whenever a patient’s experience depends on empathy, trust, and understanding, AI will come up short. People seeking treatment for addiction or mental health challenges want to speak with someone who knows what they’ve been through. No algorithm replaces that.
There’s also a boundary for AI within the admissions process. For some facilities, AI systems currently handle the early stages of admissions, such as answering program-related questions, collecting insurance or payment information, and entering data into a CRM. But AI agents then hand off a client call to a human before the intake process wraps up. That’s where it ends, at least at this point. As AI evolves, we may see that boundaries are pushed a bit further.
The other area where AI falls short is less about technology and more about the people using it. This may seem obvious, but if a worker tries to make AI do something that it’s not meant to do, the assignment will fail. For example, AI might be asked to assist professional judgment, but it can’t play a role in high-stakes clinical decisions. In this case, AI isn’t failing; it doesn’t have the capacity for accountability or ethical responsibility.
Compliance Requires AI to Mirror Humans
For behavioral health, HIPAA governs how owners and staff handle patient information, but the ethical obligations go further. If a behavioral health organization isn’t strict about setting rules for an AI agent interacting with potential patients, there will be liabilities. Medical ethical rules are still trying to catch up with this new world of AI. But owners can apply old rules: anything a human wouldn’t say, don’t let an AI bot say.
Perhaps the most critical scenario is a caller in an acute crisis. Behavioral health organizations must equip an AI system that handles initial intake calls to recognize warning signs and immediately redirect the caller, exactly as a trained staff member would. The bottom line for organizations evaluating AI vendors is this: choose systems built with behavioral health compliance in mind, not general-purpose tools that entrepreneurs have retrofitted.
Getting Staff on Board
Even after an organization successfully implements AI, the technology will fail if staff resist or misuse it. Change management is as important as finding the right technology. Staff may resist adopting any new technology for generational, technical, or philosophical reasons. Or they simply may be unfamiliar with what AI can do for them. A behavioral health organization can overcome resistance to AI by running seminars, funding staff training, and developing guidelines for what employees should and shouldn’t tackle.
To get staff to use AI regularly and efficiently, an owner must lay out a unified, purpose-driven strategy (e.g., eliminating missed calls to the call center). The organization’s owner has to set the standard for how staff will use AI.
Looking Ahead
Five years from now, behavioral health software may be unrecognizable. The data entry that consumes so much staff time (e.g., entering patient information, logging calls, updating records) will likely become automated. The way users interact with systems will shift from manual navigation to conversational interfaces.
For example, phone calls from potential clients to behavioral health organizations will tap AI to automatically populate information into whatever software system staff use, whether it is a CRM, an EMR, or something else. When staff need to retrieve information from these systems, they will start a conversation with the software, a capability some behavioral health CRMs already have.
The human connection, which is a linchpin of behavioral health, isn’t going anywhere, though. Technology, like AI agents, will clear the administrative underbrush. And by doing that, clinicians, admissions staff, and alumni coordinators can focus on helping people get well and stay well.
David Farache is CEO and co-founder of Boca Raton, Fla.-based Dazos, a provider of software to accelerate admissions, recover lost revenue, and streamline operations for behavioral health facility executives and staff. Prior to Dazos, Farache, with Louis Devaleix, successfully operated a multi-state behavioral health organization on the U.S. East Coast.
