Why Telehealth Therapy Is Vulnerable to AI
and In-Person Isn’t
This is not about which model of care is better. It is about which one occupies the same channel as AI and which one does not. The answer has real implications for how you build your career as an OT, SLP, or PT.
The pediatric therapy profession is in the early stages of a genuine reckoning with AI. Most of the conversation so far has been reassuring: AI cannot replace a skilled clinician, the therapeutic relationship matters, human judgment is irreplaceable. All of that is true.
But it is not the full picture. The risk to any given clinician’s career is not uniform. It depends heavily on how they practice. And the most important variable is one the profession has not talked about clearly yet: whether the work happens through a screen or in a room.
The channel problem
When a telehealth occupational therapist, speech-language pathologist, or physical therapist delivers a session, the experience on the child’s end is: a screen with an adult on it. The therapist is one option among all the things a screen offers. A game. A video. A show. The child knows, even if they cannot articulate it, that the screen contains more engaging things than the adult asking them to repeat a word or trace a shape.
This is not a criticism of telehealth clinicians. It is a structural observation about the medium. And it matters because AI coaching tools for children are being built inside the same medium.
An AI-driven therapeutic coaching tool for a child with a speech delay does not look like a robot doctor. It looks like an engaging app. It responds instantly. It adapts in real time to what the child does. It never gets tired, never has a bad day, never needs the child to wait while it pulls up the next activity. For a child whose primary relationship with screens is games and videos, the gap between “engaging AI app” and “adult on a video call” is narrower than anyone in the profession wants to acknowledge.
An AI coaching tool and a telehealth therapist compete in the same channel. An in-home therapist does not compete in that channel at all.
What the channel looks like for in-person care
Now think about what happens when an occupational therapist arrives at a child’s home.
There is no screen. There is a person, in the room, with a body and a presence and the ability to touch. The therapist can feel how a child’s muscle tone changes when they transition to a new activity. They can position themselves at the child’s eye level without asking anyone to adjust a camera. They can hand the child an object, guide their hand, catch them when they lose balance. They can read the room: what the family dynamic is like today, whether the dog being in the space is helping or hurting, whether the child ate lunch.
No AI tool can do any of that. Not because AI is not sophisticated, but because doing those things requires a physical presence in a physical space. The channel itself is protected. There is no screen-based substitute for a human body in the room.
Touch is not digitizable
An OT guiding a child’s hand through a fine motor task, an SLP placing a child’s fingers on their own throat to feel voicing, a PT providing joint compression for proprioceptive input. These require a body in the room. No camera resolution closes this gap.
Presence changes the nervous system
Co-regulation is not a metaphor. A calm, regulated adult nervous system next to a dysregulated child produces a measurable physiological effect. This is proximity-dependent. A screen transmits information. It does not transmit presence.
The environment is the treatment
In sensory integration, motor learning, and functional communication, the goal is performance in the real environment. In-home OT, SLP, and PT happens inside that environment. Generalization is not a separate step — it is built into every session.
Engagement is not a design problem
A child in a room with a skilled therapist who has brought the right materials, read their state correctly, and structured the activity to sit just inside their window of tolerance is engaged because the environment demands engagement. An AI tool has to compete with everything else on the screen for that same attention.
This is not about which model is “better”
Telehealth therapy expanded access to services for families in rural areas, families without transportation, and families whose children cannot tolerate transitions to clinic settings. That is real and it matters. Many skilled OTs, SLPs, and PTs built meaningful telehealth practices during the pandemic and continued because the model served their patients well.
The point here is not that telehealth therapy is bad or that in-person therapy is the only valid approach. The point is narrower: from a career durability standpoint, these two modes of practice do not carry the same AI risk profile. One competes in a channel where AI is improving rapidly. The other operates in a channel where AI has no foothold.
Telehealth OT, SLP, PT
- Delivery channel is a screen
- Child engagement competes with other screen content
- AI coaching apps occupy the same medium
- Generalization requires separate effort
- Physical guidance not possible
- Co-regulation attenuated by distance
In-home OT, SLP, PT
- Delivery channel is physical presence
- No screen-based competition for attention
- AI has no equivalent in-person tool
- Generalization built into every session
- Full physical guidance available
- Direct co-regulation with the child
Where the risk is actually coming from
AI tools aimed at pediatric developmental differences are not theoretical. Companies are building apps for articulation practice, language modeling, AAC support, and fine motor skill-building through tablet interfaces. The current versions are not clinical-grade. They do not replace an SLP, OT, or PT. But they are improving, and the improvement curve for software products is not linear.
The families most likely to trial these tools first are the same families currently on telehealth waiting lists or using telehealth as their primary access point. Not because telehealth is low-quality, but because those families are already comfortable with screen-based care and already looking for more access to support than the system can provide.
No one is predicting that AI replaces occupational therapists, speech-language pathologists, or physical therapists in the next five years. The profession is not at risk of disappearing. But the distribution of where the work comes from is likely to shift. Telehealth volume is the most plausible displacement target. In-home, in-person care is not.
Clinicians who build their practice around in-home, in-person delivery are building in the part of the market where AI does not compete. That is a durable position regardless of how the technology evolves.
What to do with this information
If you are an OT, SLP, or PT thinking about your practice over the next ten years, this is worth sitting with. The question is not whether you are good at your job — it is whether the channel you are working in is one where your value can be replicated by a better software product.
In-home pediatric therapy is not a trend or a niche. It is the form of practice that is structurally insulated from the displacement pressure building in screen-based care. The clinicians who recognize that early are the ones who will feel this shift the least.
The question is not whether you are good at your job. It is whether the channel you work in can be replicated by a better software product.
This post is part of a series on AI and pediatric therapy careers. Start with the pillar piece: Your Career Is AI-Proof. Here Is Why.
Practice in the channel AI can’t reach
Coral Care is a pediatric in-home OT, SLP, and PT network across nine states. If you are thinking about where to build your practice, come talk to us.
Learn about joining Coral Care