Artificial intelligence is rapidly transforming healthcare, but few sectors face higher stakes than mental health. While many companies are racing to build AI-powered therapy tools, Klarify is taking a different approach: using AI to support therapists rather than replace them.

Backed by Y Combinator and launching publicly after an extended beta period, Klarify has built an AI operating system designed specifically for mental health professionals. The platform helps therapists manage documentation, insurance claims, compliance, billing, treatment planning, and other administrative work that often consumes a significant portion of their time. Today, thousands of therapists across North America use Klarify to streamline their practices while maintaining full clinical control over patient care.

Moody Abdul, Founder and CEO of Klarify, is also the host of The Future of Therapy Podcast, where he speaks with clinicians, practice owners, researchers, and technology leaders about the forces reshaping mental healthcare.

I recently sat down with Abdul to discuss why he believes AI should never replace the therapist, the growing "AI arms race" between providers and insurance companies, the operational challenges facing mental health professionals, and how technology can improve access to care without compromising the human relationships at the heart of therapy.

Many people assume AI will eventually replace parts of therapy itself. You've argued the opposite. Why do you believe AI belongs around the therapist rather than between the therapist and the client?

Mental illness ran through my Lebanese refugee family, mostly unspoken. Therapy changed my life. It repaired a long, broken relationship with my father and taught me how to manage my own mental health.

That is why I reject the idea that AI should replace therapists. AI can challenge a thought or guide someone through a framework, but it cannot create the safety that allows people to say what they have never said out loud. Research consistently shows the strongest predictor of successful therapy is the therapeutic alliance between client and therapist, not the specific technique being used.

The opportunity for AI is around the therapist, not between the therapist and the client. Most of a therapist's week is spent on notes, billing, reports, and administration. Remove that burden, and you create more time for human connection.

That is also why therapist chatbots concern me. They can sound empathetic without actually understanding suffering, and they often reinforce harmful thinking rather than challenge it. Klara handles everything around the session so the therapist can remain the part that heals.

You often describe mental health providers as being caught in an "AI arms race" with insurance companies. What does that look like in practice, and why is it becoming such a serious issue for therapists today?

Most therapists rely on insurance reimbursements, and insurers have invested heavily in technology designed to find reasons to deny claims. Therapists are left fighting those decisions alone.

That creates an AI arms race. One side has software built to say no. The other side often has nothing. We give therapists AI that files claims and fights denials on their behalf.

Federal CMS data shows marketplace plans deny roughly one in five in-network claims. Almost nobody appeals, even though about 44% of appeals succeed. For a full-time therapist, that can mean roughly $12,000 in lost annual revenue. Much of that money could be recovered if someone had time to fight for it. We are building the AI that does.

What surprised you most when you began speaking with therapists about how they actually spend their time and money running a practice?

I came from the meetings software world and originally thought therapists spent their days listening to clients. My own therapist showed me otherwise. She spent nearly 40% of her time on notes, insurance paperwork, and reports, much of it unpaid.

The deeper surprise was that only about a third of a therapist's time is actual therapy. The rest is running a small business. Unlike dentists or lawyers, most therapists do it alone while carrying the emotional weight of helping people through difficult moments.

They chose this profession to help others, not manage a back office. That gap is why Klarify exists.

You've said many therapists are spending tens of thousands of dollars annually on fragmented software, billing services, compliance tools, and administrative support. How did the industry end up with so much operational complexity?

The complexity accumulated one tool at a time. Therapists adopted separate products for scheduling, notes, billing, bookkeeping, marketing, insurance, payments, and client acquisition because each solved a specific problem.

The result is that many solo practitioners spend close to $30,000 annually running their practices. Most underestimate the cost because they only see the software subscription, not the billing fees, bookkeeping, directories, marketing expenses, payment processing, and supervision costs layered underneath.

AI changes that equation. For the first time, one system can understand the entire practice and handle many of those functions together. Our goal is to consolidate that fragmented stack into a single operating system for everything except therapy itself.

As AI adoption accelerates across healthcare, what are the biggest mistakes technology companies are making when they try to build products for mental health professionals?

Technology companies make three common mistakes.

First, they try to replace therapists instead of supporting them. Mental healthcare is not a problem that should be automated away.

Second, they build generic tools that understand everyone and therefore understand no one. Therapists need systems designed around how they actually work.

Third, they treat trust as marketing. In mental healthcare, privacy cannot be a slogan. If a company says it will not train on client data, that should be a contractual commitment.

Klarify now supports thousands of therapists across North America. What patterns are you seeing in how independent practitioners and larger clinics are adapting to AI-driven workflows?

Klarify now supports more than 8,000 therapists, and about 88% of paying users remain with us after a year.

Therapists are understandably cautious about AI because confidentiality is central to their profession. Adoption is driven less by features than by trust. When clinicians see that privacy and confidentiality were built into the product from the start, they become much more comfortable.

Solo practitioners often arrive through word of mouth. Clinics usually evaluate the economics. What surprises both groups is that Klara becomes far more than a note-taking tool. About 71% of usage is for tasks beyond notes, including letters, treatment plans, supervision prep, and case support.

The common outcome is that therapists feel they get their time and enjoyment of the profession back.

There is growing concern that healthcare is becoming increasingly automated, especially on the insurance side. Do you think AI is helping or hurting access to mental healthcare today?

AI can either help or hurt access to mental healthcare depending on how it is used.

The American Psychological Association reports that more than half of U.S. psychologists cannot accept new clients. Replacing therapists with AI risks increasing isolation. Reducing administrative work gives therapists more capacity and allows more people to receive care.

The larger barrier today is friction. Insurers deny roughly one in five claims, appeals are difficult, and people often drop out of treatment. Those costs do not disappear; they simply reappear later in emergency rooms and inpatient settings.

When AI helps therapists complete paperwork, submit claims, and get paid faster, access improves. The technology should support care, not replace it.

Privacy is a major concern whenever AI intersects with healthcare. How do you approach data protection, compliance, and trust in a field where confidentiality is foundational to the profession?

The future of mental healthcare will involve AI in nearly every tool therapists use. The real challenge is ensuring privacy and trust remain intact.

We are HIPAA compliant and meet equivalent standards in Canada and the UK. We never train AI models on therapist notes or client data, and we put that commitment directly into our terms of service.

We also go beyond regulatory requirements. Even anonymized data can often be re-identified, so we do not use therapist data for model training at all. Therapy changed my life, and we build privacy standards we would want for ourselves.

Most importantly, the therapist stays in control. Klara can suggest and automate, but the clinician approves every clinical output.

If current trends continue, what does the typical therapist's practice look like five years from now? Which responsibilities disappear, and which become even more important?

Five years from now, most administrative work will happen automatically in the background. Notes, billing, letters, and marketing will consume far less of a therapist's time.

Therapists will still make every clinical decision. AI can accelerate understanding, but it cannot replace judgment. The human relationship becomes even more important because it is the one thing technology cannot replicate.

I also expect therapy to extend beyond the weekly session. AI can help maintain momentum between appointments through guided reflections and therapist-directed support. The goal is not to replace therapy but to make the time between sessions more meaningful.

What disappears is administration. What becomes more valuable is presence, judgment, and trust.

Stepping back from Klarify specifically, what do you think the mental healthcare system gets wrong about supporting therapists, and how could technology help fix those structural problems without compromising patient care?

The system asks therapists to be clinicians, business owners, marketers, billers, and insurance specialists all at once. Then it penalizes them for paperwork mistakes and wonders why burnout is so common.

Two changes would make a major difference. First, claims should be simple to submit and fast to pay. AI can help make that possible. Second, administrative work should be removed from the therapist's workload entirely.

The bottleneck in mental healthcare is not just the number of clinicians. It is the amount of their time consumed by work that is not therapy.

The line that cannot be crossed is care itself. Therapists must remain responsible for every clinical decision. Technology should carry the load around the work, never become the work.

Brendan Blowers is a writer and journalist whose work has appeared in Haute Living, International Business Times, TechTimes, Haute Residence, TOWN, Vocal, and elsewhere. His work spans culture, lifestyle, technology, business, architecture, creativity, and leadership.