What 1 Year of Building an AI Training Platform Taught Us About What Therapists Actually Need

When we started building Tmind AI, we thought we were creating a simple therapy training tool.

One year later, it’s clear:
what therapists actually need from training tools is not what we expected.

But before we get into that, it’s important to zoom out.

The conversation around mental health access often focuses on one thing:
we need MORE therapists.

That’s true, but it’s only part of the picture.

• Only 27.3% of mental health needs in the U.S. are currently met1

• Over 30,000 psychiatric positions remain unfilled2

• 137 million Americans live in mental health provider shortage areas3

• The U.S. Bureau of Labor Statistics projects 18% growth in demand for mental health counselors through 20324

This points to a clear reality: demand is accelerating faster than supply.

But there’s a second bottleneck that gets less attention: The clinical training system itself is already at capacity. Even as programs try to enroll more students, training still depends on:

• Limited role-play hours

• Overextended supervisors

• Practice environments that are difficult to scale

What tools can expand training capacity without increasing faculty burden?

That’s the question that led us to build Tmind AI, and the lens through which we’ve interpreted everything we’ve learned.

We initially believed students needed more practice. They DO, but the deeper issue isn’t just skill development. It’s clinical confidence.

Professor at Salisbury University shared:

“I’ve seen a real boost in my students’ confidence. The vast majority felt proud and excited as they saw their skills improving.”

This wasn’t about scores or technical proficiency. It was about a shift:

• From “I understand the theory”

• To “I trust myself in the room with a client”

That shift is what determines readiness for real-world therapy.

And it only happens when practice feels real enough to carry emotional weight.

We originally positioned Tmind AI as an additional training resource. But the feedback from counselor education programs was consistent:

“We don’t need more tools. We need less burden.”

Faculty are balancing:

• Teaching full course loads

• Supervising practicum students

• Managing accreditation requirements

• Maintaining clinical responsibilities

The features that actually drove adoption weren’t the most advanced, they were the most practical:

• Automated evaluation templates that generate feedback

• Progress dashboards that reduce manual review time

• Custom AI clients aligned with existing course materials

One of the most overlooked barriers in clinical training is emotional: the fear of failing in front of others.

In-class role-play is valuable, but it can also create pressure that limits learning.

We heard from a student at York University who practiced CBT techniques using Tmind AI. Before entering practicum, he said he finally felt: “like a therapist.”

Not because the AI replaced training, but because it created space to:

• Try

• Make mistakes

• Reflect

• Improve

Without being observed or judged.

Just the ability to learn without pressure.

One year in, the core question guiding our product has changed.
We used to ask: “How do we make the AI more realistic?”
Now we ask: “What does this student need to feel ready for their first real client?”

That shift changes everything.

It moves the focus from:

• Technology → to education

• Features → to outcomes

• Simulation → to preparedness

And the answers are often simpler than expected:

• Timely, actionable feedback

• Practice scenarios aligned with coursework

• Visibility for supervisors without requiring more time

The shortage of mental health providers is not just a pipeline issue.

It’s a training infrastructure problem.

If we want to prepare more therapists, without overwhelming educators, we need:

• Scalable practice environments

• More efficient feedback systems

• Tools that extend, rather than replace, existing training models

This is bigger than any single platform. It’s a systems challenge that requires new approaches across the field.

We’re still early. One year is just the beginning. But one thing is clear: The future of therapist training isn’t just about better AI. It’s about better preparation.

If you’re a faculty member, supervisor, or student, we’d love to hear your perspective.

What’s working? What’s missing?

Because this isn’t a finished product. It’s an ongoing conversation. 💚

Leave a comment