Tony Lee, CTO at SimplePractice – Interview Series

Tony Lee, Chief Technology Officer at SimplePractice, is a veteran technology executive with more than two decades of experience building products, scaling engineering organizations, and modernizing complex technology infrastructure. Before joining SimplePractice in 2024, he served as CTO of Hyperscience, where he oversaw product delivery and technology strategy for an enterprise machine learning automation platform. His previous leadership roles include Head of Engineering, Security and IT at Snapdocs, VP of Engineering at Dropbox (DBX ) and Zendesk, and senior engineering positions at Bitcasa, Box, and Yahoo, where he led the engineering organization behind Yahoo! Finance. Lee began his career working on enterprise software at Vitria and wide-area network research within NASA’s Advanced Supercomputing Division.

SimplePractice provides electronic health record and practice management software designed for therapists and other health and wellness professionals, including psychiatrists, occupational therapists, and speech-language pathologists. Its platform brings together scheduling, billing, clinical documentation, insurance tools, secure communications, ePrescribing, client management, and integrated telehealth to help independent and group practices manage both care delivery and administrative work. SimplePractice reports that its platform is trusted by more than 250,000 practitioners serving over 20 million clients.

You spent decades leading engineering organizations at companies such as Yahoo!, Box, Zendesk, Dropbox, Hyperscience, and now SimplePractice. How has your perspective on building mission-critical technology evolved, and what lessons from those experiences have been most valuable in developing AI tools for mental healthcare?

This moment in time is so interesting as it’s driven by advances in LLM-based technology. In our industry, the temptation is always to treat the technology itself as the product. But as a customer-focused SaaS company, it’s not. Our role is to deliver value to our customers, specifically to help independent mental health practitioners thrive and retain their independence by simplifying the business of running a private practice.

This doesn’t mean that we shouldn’t invest in technology innovation; but, it should always be with a focus, even if it’s on an undefined timeline, on how this will eventually help customers, empowering them to stay focused on care by reducing their administrative burden.

SimplePractice recently launched an AI-powered clinical assistant developed with input from more than 1,000 therapists. What did that collaboration reveal about what clinicians actually want from AI, and were there any assumptions your team had to rethink during development?

Care Aide is SimplePractice’s new AI-powered clinical workflow assistant designed to help independent practitioners reduce administrative burden and stay focused on care. This includes Note Taker, our AI clinical transcription tool, plus new tools to support clinicians across session preparation, documentation, treatment planning, intake review, and client summaries.

When building Care Aide, we incorporated feedback from more than 1,000 mental health practitioners so that our technology aligned with real-world clinical workflows. From this collaboration, we heard clearly that clinicians don’t want technology to be a decision-maker, but rather a tool to support administrative work. We developed this suite of tools with this measured implementation in mind to ensure that we’re keeping clinicians in control, while helping with practical work that leads to administrative burnout.

Mental healthcare presents unique challenges compared to many other industries adopting AI. What design principles guided your team to ensure the technology supports clinicians without influencing or overriding their professional judgment?

It’s important to remember that our role is to help clinicians, who must always remain in the center and in control. With that in mind, we focus on keeping humans in the loop by design, being transparent where AI is being applied, and giving our customers control over its application. I believe that these principles are not only incredibly important in mental healthcare, but are broadly applicable across industries.

You have previously spoken about the importance of human-in-the-loop systems in enterprise AI. How does that philosophy translate to therapy settings, where trust, nuance, and clinical expertise are central to patient outcomes?             

While human-in-the-loop remains important for all enterprise AI, it is especially important in therapy settings where clinicians remain at the center of patient care. AI should be seen as a tool to aid administrative tasks, allowing practitioners to spend more time on patients and direct care delivery.

Our AI roadmap is purpose-built on a foundation of security and trust; SimplePractice is HIPAA-compliant and HITRUST-certified, and our AI tools are built to align with and be even more protective than HIPAA requirements. Clinicians always remain in control with reviewing, editing, and approving outputs before they become part of any care documentation or are shared with clients.

As AI becomes more deeply integrated into healthcare workflows, what technical frameworks are necessary to maintain patient privacy, data security, and regulatory compliance at scale?

We know that the threat landscape has changed dramatically with AI. Externally, vulnerability discovery and exploitation are moving much faster, and internally, Agentic AI will pressure test internal controls as the drive to connect internal systems grows.

This increases the importance of security by design as well as friction to the access of critical data even at the cost of AI efficiency. I am all for AI-based automation and have worked on this across multiple companies, but in order to use AI with confidence, we need to know we can use it safely.

How do you approach balancing automation with clinician autonomy, particularly when there may be pressure to automate more aspects of documentation, note-taking, or patient interactions?

At SimplePractice, our approach is to provide the tools and capabilities, while respecting the clinician’s choice on how to use them. Because our customers are primarily independent private practices and they chose this route so they can practice in the way that works best for them and their clients, the balance on how they implement these tools into their day-to-day workflow is entirely up to them. We want to show them the ‘art of the possible’ as technology is evolving so quickly, but they stay in control.

What role does real-time clinician feedback play in the development process, and how do you ensure practitioner insights are incorporated into engineering and product decisions?

Real-time clinician feedback is critical to the product development process, especially when it comes to supporting clinicians in delivering care to patients. At SimplePractice, we have multiple layers to this approach. First, we have an internal Clinical Strategy team that includes licensed clinicians who work closely with our product team from ideation to development. We also have an external Clinical Advisory Board as an additional early source of feedback to guide product innovation and ensure clinical accuracy, and a Customer Advisory Board, SimpleCircle, to gather direct insight on platform experience and to keep our tools aligned with the rapidly evolving needs of customers.

Once the idea has started to take shape, we typically use a series of user feedback sessions through long running early adopter programs to get increasingly detailed feedback. Finally, we have very active and candid community forums, as well as host our own livestreams to connect with our customers to understand what parts of our platform are working well and how we can continue to innovate to best serve independent practitioners.

Many healthcare professionals remain cautious about AI adoption. What concerns do you hear most often from therapists, and how is SimplePractice addressing those concerns through both product design and transparency?

Caution around AI adoption is completely understandable. We strongly believe that clinicians must have a choice in whether they use AI in clinical care.

But choice only matters if clinicians can trust the tools available to them. SimplePractice is HIPAA-compliant and HITRUST-certified and we recently launched the SimplePractice Trust Center to give full visibility into our security standards and data practices.

We believe clients deserve to know when AI is part of their care and provide consent and disclosure tools so clinicians can inform clients before AI is used, helping them meet evolving state requirements while keeping the therapeutic relationship grounded in trust.

Our AI tools keep clinicians in control. AI-generated content is always a draft, and clinicians review, edit, approve, or disregard it. They decide what belongs in the clinical record and remain accountable.

Ultimately, our job is to earn trust over time by building tools that work safely, transparently, and in partnership with clinicians.

Administrative work is a major source of burnout across healthcare. Which tasks are most appropriate for AI to handle today, and where do you believe human involvement should remain indispensable? 

Today, AI can help with handling the repetitive, organizational tasks that can lead to administrative burden. In our Annual State of Private Practice report (2025 in review), we found that uses for AI in clinical work included treatment planning, writing emails, and marketing. Clinical documentation is another helpful area that AI can support. For example, our Note Taker tool saves clinicians up to 5 hours per week on average. However, human involvement in mental healthcare remains indispensable, especially when it comes to clinical judgment and the therapeutic relationship. AI should never substitute for a clinician’s expertise. The technology is there to handle the logistical weight so that the human relationship can remain the central focus of care.

Looking ahead, how do you see AI reshaping mental healthcare over the next five years, and what safeguards must the industry establish now to ensure these technologies enhance care rather than compromise it?

Therapy is about human-to-human connection and the power lies in the therapeutic alliance between the clinician and client. AI has the opportunity to strengthen this in many ways, but the biggest near-term needle mover is the gift of time. By applying AI to help with all of the administrative tasks and worries a clinician has outside of the actual session – ranging from dealing with insurance and billing, to managing a very busy and dynamic calendar, to preparing for a session by reviewing notes and the latest research – AI can reduce the time spent on all of these non-clinical activities while at the same time doing it as well as, and in many cases, better than the clinician. As a result, clinicians have time back to see clients and think about treatment planning. This is the immediate promise of AI in mental healthcare that I see for the clinician: time to focus on the human-to-human connection that is so central to delivering care.

Thank you for the great interview, readers who wish to learn more should visit SimplePractice.