Blog

The Hardest Yes in Health Care: What AI Has to Get Right for Behavioral Health

Written by Anonymous Health | Oct 14, 2026, 12:04:17 AM

It took them months to reach out. It should take seconds to answer.

Most of health care runs on a referral. Your doctor finds the problem, orders the procedure, and sends the order to the hospital. By the time you call to schedule, the decision has already been made. All you have to do is show up.

Behavioral health does not work that way. The person usually diagnoses themselves, searches for a provider on their own, and then picks up the phone to make one of the biggest health decisions of their life. They make it sight unseen, often with no one they trust to tell them good care from bad.

Bob Poznanovich, who has run admissions teams and helped shape our HeadsUp product, calls this the hardest yes in health care. The numbers he shares when he speaks make the point. About 80% of people with a chronic illness like diabetes get diagnosed and get help. For mental health, it is fewer than half. For substance use, it is about one in ten.

It is not one yes. It is five.

Bob describes the path into care as five separate decisions, and a person can stall at any of them:

  1. Do I have a problem?

  2. If I do, what do I do about it?

  3. Where do I go?

  4. Is this the right place for me? Do they take my insurance, do they have room, did they make me feel comfortable?

  5. Will I actually show up?

Somewhere along that path is what Bob calls the moment of clarity: the point when a person decides that now is the time to ask for help. That moment closes quickly. A voicemail can close it. So can a long hold, or a benefits check that will not come back until tomorrow. When it closes, the next opening usually arrives with the next crisis, and no one controls what that crisis will cost the person, their family, or their workplace.

Bob used to tell his phone teams that their job was to be present for the call and build a relationship, because they might be as close as that caller ever gets to care.

That is the standard any AI in this field has to meet. Here is what it takes.

1. Open more doors, and be ready when they are

Not everyone is ready to talk. Chat gives someone a first step they would not take by phone, as long as it actually answers the question instead of handing back canned responses. As Bob puts it, chat reaches the person who is struggling, while the phone more often reaches the caregiver.

No team will answer every call, and voicemail is where intent goes to wait. Few people leave a message, and those who do are hard to reach again. AI should make sure a missed call is not a missed person. You can't answer every call. But what if you could?

2. Listen before it asks

When Bob ran an admissions team, his staff worked with three screens: the CRM that sales and marketing wanted, the health record the clinical team needed, and the phone system flashing who was on hold. Each one came with its own list of questions.

He recently listened to a call from a mother whose son had overdosed and was intubated. The first thing she heard was a question about her insurance. It was the next field on the screen. It was not the next thing one person should say to another.

AI that supports a clinic has to follow the conversation, not the form. Its job is to help the person on the phone find the right next thing to say.

3. Know where the person is in their journey

Someone who is still deciding whether this is the right place cannot answer questions about admission. HeadsUp was built on the science of motivation and change for that reason. It listens for where a caller is on the path, then offers the guidance that fits that moment, including when it is right to ask for the next commitment.

Done well, the caller feels like a partner in their care, not someone being pushed into it.

4. Bring every person on the team up to the level of your best

Bob saw it when he ran sites. Same team, same office, same phone system, calls arriving at random, and some people bring twice as many callers into care as others. New staff are the most eager to pick up and the least experienced. Shifts vary. Training varies too, and much of it is about systems instead of conversations. The success of the call depends on who picks it up.

Referral sources notice. Bob stopped sending people to facilities whose phones left a bad impression, because the referral carried his name.

The answer is not to replace people. Most calls, and every admission, still happen live with a human being. AI earns its place by coming alongside the team so that your best admissions conversation happens on every call, every chat, and every shift. It is not about finding fault with anyone. It is about understanding the system they are working in.

5. Think past the admission

Hospitals are paid by the episode. Behavioral health runs on patient days and patient sessions, and outcomes improve the longer someone stays engaged in care. Bob's math is simple: a patient who stays 3% longer has the same effect as 3% more admissions.

So the first conversation has to do more than fill a bed. What a clinic learns at intake should travel with the patient, so the care team can meet them where they are and keep them engaged long enough for treatment to work. The first conversation is the front door.

6. Speak health care, not sales

The people answering these phones are trying to keep patients alive. Technology that talks about them as leads in a pipeline does not belong in the room.

Mission and margin are not opposites here. More people in care makes a program more sustainable. A sustainable program can keep its doors open and reach communities that have gone without. And if care does not exist, neither does hope.

Where to start

It takes courage to reach out. It should not take luck to get an answer.

If you want to see how your own front door is working, ask us about a complimentary Front Door Assessment. We look at the system, not the people in it.