AI is genuinely useful in the work of finding the right educational and therapeutic placements for young people. It is also missing the parts that matter most.
My daughter lives in San Francisco and works inside the world where AI is being built and debated and reimagined daily. When she visited recently, the conversation turned (as it tends to when she is home) to where technology is heading. I told her what Rebekah Jordan and I have been exploring at Crossbridge Consulting: how we use AI in our practice, how it helps us synthesize research, organize complex information, and move faster on things that used to take hours. She nodded. Then she said something I have not stopped thinking about: most of what has been written about AI is wrong, and you really have to know how to use it.
She is right. And that is coming from someone who knows.
What she meant is this. Most of the content about AI was written before people understood it well, by people who still do not, and AI is learning from that content. It is, in a very real sense, trained on its own misinformation. The same problem exists when families turn to AI to research therapeutic programs or independent schools. The technology surfaces whatever has been written, with no ability to evaluate whether what was written is accurate, current, or complete. The clinical director who left six months ago. The program culture that has quietly shifted. The peer group that bears no resemblance to what is described online. Garbage in, garbage out, except the garbage arrives dressed in confident, well-organized prose, and there is nothing in the output that tells you not to trust it.
For families making high-stakes decisions about their children’s care, that is not a minor limitation. It is the limitation.
I want to be careful here, because reflexive skepticism about AI is as unhelpful as uncritical enthusiasm. AI is a remarkable technology. In the right hands, with the right questions, it genuinely extends what a knowledgeable practitioner can do. It synthesizes research across hundreds of sources in seconds. It helps a frightened parent at 2:00 a.m. understand what a diagnosis means, or what questions to bring to an appointment. It helps clinicians draft documentation, organize case notes, and identify patterns across large amounts of information. We use it. We are grateful for it.
And then there is everything it cannot do.
What Walking the Halls Reveals
I was recently in California visiting Key Healthcare, a program serving teenagers navigating some of the hardest seasons of their young lives. I sat down with the founder, Ryan Blivas, and he told me something that has stayed with me. Ryan had been sent to treatment himself as a child. What shaped his entire vision for Key Healthcare (the thing that drives every decision he makes about how his program runs) is that he knows what it feels like not to be seen. Not to be heard. He built something specifically designed to make sure his clients never feel that way.
No algorithm surfaced that for me. No database contains it. I was sitting across from him, listening, and I felt the weight of it: the way a founder’s personal history becomes the soul of a program, for better or worse. In this case, profoundly for better.
This is what I do when I evaluate a program for a family. I am not reading a brochure. I am walking the halls. I am checking behind the toilets and pulling back the shower curtains (yes, I do that), because mold and neglect tell you something marketing materials never will. I am asking staff what they love about working there and watching their faces when they answer. I am watching the kids. Are they engaged or disengaged? Bright-eyed or dulled out? You can feel whether kids in a program are living in it or just surviving it. You can feel whether the people running it genuinely care about their work, about these particular teenagers, about what happens to them after they leave.
Nobody writes about that. You cannot write about it without being there. And you cannot be there meaningfully without knowing what you are looking for.
That work requires a person who has spent years learning how to see. And then it requires them to show up and look.
The Cost of the Wrong Fit
A therapist called me not long ago with a referral. She had been working with a young man (bright, complicated, wonderfully quirky, and struggling with substances) whose parents had tried twice to find the right therapeutic program for him on their own. Both times, they had turned to ChatGPT. Both times, it had given them something that looked like an answer: a program name, a description, a sense of direction when they were desperate for one.
Both placements were wrong fits. He was discharged from both.
I want to sit with that for a moment, because we do not talk enough about what a failed placement actually costs a kid. It is not just logistical disruption, though it is that too: the frantic calls, the scramble for next steps, the financial strain of starting over. It is what a discharge does to a young person who is already struggling to believe in himself. Each time this young man left a program that was not right for him, he did not think the program was wrong. He thought he had failed. Again.
That is the cost of the wrong fit. And it is a cost that does not show up in any search result.
His therapist, who understood what this family actually needed, insisted they hire an independent educational consultant. Not a search engine. Not an AI. Someone who would know the programs, not from a database but from having walked through them, met their clinical teams, understood the specific culture and approach and peer population each one actually serves. Someone who could hold this particular young man (his quirkiness, his substance use, his specific history of what had not worked) and find the place built for exactly him.
That is the call I got. And that is the work I do.
The Golden Thread
There is something else a human consultant brings that no search can replicate. Knowledge of a client deepens over time. I stay connected to the therapist, hear about the trials and tribulations and emerging strengths, understand how the child is responding to structure and where they are ready to have the training wheels pulled back (metaphorically speaking). I know which programs provide more structure and which provide less, and I know when a particular child is ready for which.
When a new program comes into the picture, the family does not go in blind. The treatment team hears: yes, they did that in the last program too. Here is what helped. Here is what did not. That history does not have to be reconstructed from scratch. It travels with the child and lands in the hands of the next treatment team as something they can actually use.
That is the golden thread running through the whole journey. The child does not have to explain themselves to every new team. Someone already knows them, and speaks for them.
The young man whose parents had used ChatGPT is now in a program that fits. His story is his, and there is nothing more to say about it here. But the therapist who made that call, who recognized that this family needed a human expert and not a better search query, made a decision that changed the trajectory of his care.
That is what knowing the difference looks like in practice.
Information Problems and Presence Problems
There is a useful distinction I have come to lean on in this work. Some of the questions families bring me are information problems. Many of them, in the end, are presence problems.
An information problem is one that better data, faster synthesis, or sharper analysis genuinely helps to solve. AI is good at these. It can help a parent understand a diagnostic framework, summarize research on a treatment modality, or compare educational approaches at a high level. It can do in seconds what used to take hours, and the seconds matter when a family is in crisis.
A presence problem is something different. It is the gap between knowing what to do and being able to do it. You can read every book on child development, understand every diagnostic framework, have access to every resource, and still freeze when your teenager is throwing things and screaming at you and you need to stay present in a way no amount of preparation quite covers. The knowing and the doing are not the same thing. And the gap between them is where most of the hard parenting, and most of the hardest decisions in this work, live.
No technology built yet closes that gap. I am not sure one ever will.
When There Is No Right Answer
I am working right now with a family whose child has schizophrenia. I will not say more than that, because their story belongs to them. But I will tell you what the question looks like. Do they choose the program that is the best clinical fit for their child (the one most equipped to serve someone with his specific and serious needs) even if it means he is far from home? Or do they choose the program that is closer, where they can see him more often, monitor his well-being more directly, be present in the ways that parents of seriously ill children need to be present?
There is no right answer. Families often come hoping there is one. What is possible instead is laying out every dimension of the decision as clearly and honestly as possible (the clinical considerations, the logistical realities, the ways each choice ripples outward into the family system) and then staying in the room while they make it.
That staying is the work. It sounds passive. It is not.
Some families come to me in the middle of a decision. Others have moved past that. They have arrived at something harder: the recognition that their child’s struggle is not a problem to be solved. It is a life to be navigated.
A mother called not long ago, in tears. Her son’s mental health struggles are not a chapter in their family’s story. They are the story, ongoing, evolving, without a foreseeable end. She had arrived at that understanding in a new way, and the grief of it was fresh. There are no perfect solutions, she said. I know that now.
She was not asking anyone to fix it. She had moved past that. She was asking for someone to think through the options with her: the trade-offs, the uncertainties, the things she could not fully know until she tried them. To do that thinking with her, not for her. From the perspective of someone who had walked this road with other families. From the perspective, she said, of a mother.
There is no AI equivalent of that phone call. There may never be.
What the Work Requires
Our profession is at an inflection point. AI is here. It is useful. It is going to keep getting better. The temptation, for families and for some practitioners, will be to reach for it as a substitute for the slower, more demanding work of human expertise. That substitution will sometimes work. And it will sometimes cost a child their next placement, their belief in themselves, their family’s confidence in the path forward.
The work of finding the right educational and therapeutic placement for a young person is not, at its core, an information problem. It is the embodied knowledge of programs you have walked through. It is the longitudinal relationship that turns a stranger into someone who already knows the child. It is the willingness to sit with parents while they make a decision that has no right answer, and to stay there while they make it.
That work requires a person who has spent years learning how to see. And then it requires them to show up and look.
By Jennifer Benson, MSW, 51³Ô¹Ï Associate (CT), Crossbridge Consulting