Good morning, everyone, and welcome to the last day of the 44th Annual J.P. Morgan Healthcare Conference. My name is Tien Wilson, a member of the healthcare group based in New York. I'm pleased to introduce Mr. Jon Cohen, who's the CEO of Talkspace, and Ian Harris, the CFO. Thank you, Jon.
Good morning, and thank you for the invitation this morning. The market remains incredibly large and unpenetrated for mental health services. If you look at the market specifically, there are 46 million Americans that have an issue with mental health services, which is 23% of the adult commercial population, 2 million military of the 10 million population in the military, 17 million, or 25% of the 66 million who have Medicare coverage, and upwards of 10 million of the 20 million teens, 50% in the ages of 13 to 17. 40% of Talkspace patients, despite this, are new to therapy. The number one requested employer benefit is mental health services, and the biggest motivator for people coming to Talkspace is the need for coverage, or actually the fact that they have coverage is the reason they come.
44% are aware of Talkspace that actually have insurance that they can use it for their benefit. With our 200 million covered lives, commercial lives, including the Medicare, and now 10 million TRICARE lives, and our teen initiatives, we remain uniquely positioned to continue to serve this growing and underserved market. Starting three years ago, about right after the prior eight months, we began our pivot to the commercial insurance part of the business, fee-for-service market, which has fueled constantly our growth over the last several years. Revenue growth is annualized rate at about 23% over that period of time, driven by payer session growth of almost 30% per year. In 2025, 2025 marks our second year of EBITDA positive performance, which at the midpoint of our guidance basically has doubled since last year from $7 million to approximately $15 million.
Our operating expenses, as a percentage of revenue, continue to go down, driving improvement in EBITDA and operating leverage. Our continued growth in the payer segment is driven by two major factors. Although we already cover over 200 million lives, we will continue to add more regional and local plans throughout the year. More importantly, our initiatives to activate these 200+ covered lives are specific to five different strategies, which I'll discuss in more detail. First, we are driving growth by increasing our consumer awareness. Remember that only 44% of people know that Talkspace is a covered service, and this includes four specific channels to drive the growth. Paid media where we target the right person at the right time with AI insights to optimize what people most likely will, or the people who will most likely need therapy.
Secondly, through organic growth, is to find more people that are utilizing LLMs for search. We have specifically designated people who have devoted their time to SEO search to optimize strategies to increase our visibility. Our marketing and payer partnerships continue to drive high-intent users to us, and our brand recognition continues to increase with our relationship with Michael Phelps and others. But most importantly, what has happened over the last several three years, actually, is our awareness continues to grow while our spending on marketing has had a significant decrease over the last three years. Second, we are driving growth by two types of partnerships that drive high-intent customers to us. One is our distribution partners, such as Amazon through their Health Conditions Program, and booking services such as Zocdoc.
Secondly, we have strategic partnerships with over 20 different entities, as mentioned here, and they refer our patients to us, as for example, our recently announced Tia for women's health, who refer to us because not only do their patients get access to Talkspace and mental health services for either no or very little out-of-pocket cost, but they also share our content each month with their entire population. Third, we are driving growth with our deepening relationship with the payers. This includes their recognition of the following: fast therapist matching and appointment, quality oversight of the provider network, which includes different quality metrics, including service quality, clinical quality, productivity, client experience, and documentation. We have defined, measured outcomes. We have surveys that reflect very positive member experience. We've had significant integration with our AI capabilities, and they're extremely happy that we are in network nationally with Medicare.
As a result, we do designated credentialing and perform credentialing for the payers. We've developed value-based contracts for them in partnership with them, and importantly, we now have full directory integration with multiple of the large payers across the system. Fourth, we are driving growth with our targeted innovations across every stage of the patient journey, which has delivered measurably significant improvements, including a 22% increase in checkouts and almost a 50% increase in what we refer to as 3 in 30, which means a patient booking three sessions within the first 30 days. So we've dissected every single step of the journey throughout the last year and includes AI integration in almost every single one of those steps, including eligibility, registration, scheduling, sessions, and in-between sessions, and billing.
As a more specific example, as you may remember, last year we announced, at this conference, Talkcast, which is our individualized, personalized podcast, which has resulted in an 8% increase in booking of second sessions and a 14% increase in booking of third sessions. We also provide now to the therapists smart evaluations on intake, which has had significant improvement in their time to devoting more time to patient quality and patient services. And we've also developed smart insights, which has resulted in a 20% increase in booking of third session and a 21% increase in third session. I'm sorry, 20% in the second and 21% in the third. Fifth, and finally, we are driving growth by expanding our offerings. Last year, I talked about, and more recently in the last several quarters, about our initiative around psychiatry. Psychiatry is mostly a medication management system.
We spoke about that throughout the year, and we now have, as a result of those initiatives, approximately 300 prescribers who are qualified to prescribe drugs, mostly anti-anxiety and mostly antidepression, and as a result of that initiative, we've seen almost a 50% growth in the last several quarters in the psychiatry business. In addition, as you know, we did an announcement with a partnership with Amazon through Amazon Rx, which helps medication adherence, and almost as importantly, we developed a mechanism to make it incredibly easy for our own therapists to refer to our own psychiatrists when patients need medication management. We also announced recently our acquisition of Wisdo. Wisdo is a lower acuity peer-to-peer and coaching platform.
It opens up particularly the area and addresses the needs of Medicare patients and seniors because their offering is particularly applicable to patients who are suffering from loneliness, who look to peer-to-peer and coaching groups to help them through those journeys. In addition, we were very proud that Wisdo announced their relationship with Novo Nordisk. What that is, is they provide a coaching journey for patients who are taking GLPs and helps patients getting through their weight management because they're particularly looking for other people who are going through the journey to help them get through, which helps, of course, renewal of the drug. Third, we've seen a significant increase in the number of patients in TRICARE and the military, particularly their families, which continues to grow, and Medicare continues to grow month over month as a result of being in both standard Medicare and Medicare Advantage.
And finally, I do want to mention our youth program. Although it's not related to the payer business, we now have over 500,000 teenagers throughout the country who have free access to Talkspace because of our contractual relationships with governments. This includes New York City, Baltimore, the City of Seattle, and recently the State of North Carolina, where we're providing teens actually who are part of the juvenile justice system access to Talkspace. In addition, multiple other private schools and multiple other charter schools. So we're particularly proud of what we've done for the teen mental health crisis across the country. So I want to pivot now to AI and what's going on relative to mental health chatbots or AI agents. And I think most of you hopefully will have seen.
Not a week goes by, essentially, where you don't read something in the press or you see some broadcast study about what's going on relative to the mental health issues and the harm that has occurred with several of the large general AI chatbots that are out there. The general purpose LLMs, which are the ones that are out there now mostly, now utilized by approximately 10% of the world's population, 800 million people or more, have been incredibly successful because they are always available. They're low cost, they're anonymous, easy to access, and essentially it makes them the default option if you want to have a mental health conversation or basically anything that's bothering you for the day. They do excel at being incredibly fluent, engaging, and very responsive to almost any type of prompt.
Unfortunately, well, let me say fortunately first, that we do believe, or I believe, that in some sense this is a very good thing relative to the democratization of mental health, meaning millions of more people now have access to some sort of support. However, despite the fact that all these millions of people have access, the general purpose LLMs were never built to support mental health, leading to what I referred to earlier as a rash of reported harmful outcomes. Mental health support requires something much more specialized and nuanced, including challenging distorted thinking, recognizing delusions, and identifying risk in real time.
So we made a decision about a year ago now to design what we refer to now as the first safe AI agent specifically designed for mental health support, developed utilizing clinically recognized standards of care, and incredibly important, its privacy HIPAA protected for all the information and discussions that a patient has. I'll talk about our database in a second. The LLM is trained and fine-tuned on Talkspace's massive mental health dataset to improve risk recognition, support appropriate decision-making, and avoid the pitfalls already seen with general purpose AI systems. It keeps clinicians in the loop constantly with clear escalation pathways to connect users to real human therapists as needed. It does not replace clinicians, but actually extends their reach. It adheres to strict clinical guidelines and standards while identifying millions of new users who will require human intervention.
I do believe that the need for human care by trained therapists will actually increase as millions more people will be identified that need professional help beyond that that could be provided by an AI agent. So I'll just mention a couple of things here relative to what it is, how to use it, and maybe what people use it for, just briefly. What is it? It can share your fears or insights at any time. It's a space to work things out and get clarity. It's a space where you confide without judgment or time limits, and it provides fast support and lasting progress. People can turn to it because it's trusted and science-backed. You can talk to it as a reality check and stop spiraling or avoid burnout, and you may want to vent with it instead of blowing up on your group chat.
It's for those people who are seeking clarity without judgment, for those searching for answers in the wrong places, and for those whose friends have stopped listening to them. Other possible use cases for the Talkspace LLM AI agent are as something to be used for guiding a patient through an interactive Q&A and matching them to the right therapy and the right therapist. It could be used to engage patients between sessions to support adherence to their care plan. It could be used to screen postpartum mothers for depression and guide their early therapy. It could be used by primary care physicians identifying mental health concerns and guiding patients to therapy.
We like to refer to it as something that is always on, always available, available 24/7 to every postpartum woman, available to every college student as needed, available to every active military, a companion for cancer patients as they go through their journey. It is what we consider the first real safe mental health LLM. And to reiterate, and I'll talk about, show you the model in a minute. It is trained on our database, one of the largest mental health databases in the country. It is, I'll reiterate, HIPAA protected and patient consented. It has within it proprietary algorithms that we've developed since actually before 2019.
These proprietary algorithms run in the background, and they actually predict and alert the therapist to suicide risk, homicide violence risk, the possibility that someone is being abused at home, substance use, and six other specific clinical entities all running in the background to identify risk. Because of those identifications, we have continuous monitoring with live professional therapists watching what goes on and all the interactions, and as needed, escalate patients out of the AI agent and into real therapy. This is a picture of what our database looks like. This is what we use to fine-tune the LLM. It's eight billion words, 140 million messages, 4.3 million psych notes, three million therapist ratings, 1.5 million treatment goals, and close to four million clinical interventions. It is a massive specific mental health database that we use to train the LLM. This is a picture of the model.
Essentially, it works like that, right at the beginning, in terms of intake. We make a decision about whether or not a patient is appropriate for an AI agent or not. So specifically, if someone comes in and says they're thinking about suicide, we're not going to put them into the AI agent. So there is a screen before you can get to the AI agent. Once you're in, people have the conversation, and as I said, it is being continuously monitored for any risk. If there's any risk, it goes off to the monitoring clinician, and then the clinician decides if we need to escalate it further to live therapy and to stop the conversation. It is currently in beta testing mode, and it will be live sometime in the first half of 2026.
So as you can see, we have come a very long way in three years in a very, very positive journey. However, there remains a tremendous opportunity in front of us, and we are positioned to continue to aggressively take advantage of that opportunity. In addition to the core business, we believe we are strategically positioned ourselves to be a leader in the application of AI to mental health services in this rapidly moving current environment. As we close out 2025, we have continued to see strong growth and profitability in Q4, and we are pleased that we ended up for the year where we have ended up for the year and continue to see strong momentum into 2026. And with that, I'll open it up to questions.
Thank you, Jon. I appreciate your remarks and the presentation. For the audience, well, if you have any questions, please, I invite you to ask. I have a few prepared questions here so we can at least start the conversation. Jon, maybe just to kind of start from the beginning of your presentation, we'll talk about covered lives. So you've mentioned that now Talkspace is a covered benefit for over 200 million people. So when you look into 2026, where do you think Talkspace can go from there?
So as I mentioned at the beginning, we will continue to add on regional plans. As you may know, there's literally hundreds and hundreds of other plans out there. We have a separate team that's set up, a commercial organization that just does payer relationships and payer contracts. So I think we'll continue to see growth, but we're at that top part of the curve where it's not as imperative to us as it was before. Our whole strategy is actually to activate the 200+ million lives we have now.
Sure, sure. And maybe turning now towards Talkspace from a segment of its peers and also looking at it from that lens, what do you think just sets you apart from your peers? And we know Talkspace is one of the only pure-play public companies. So just what makes you different?
That's a great question. So I think one of the most important differences for us is we have, or we refer to as, and I talked a little bit about the quality oversight, a real curated network. And what I mean by that is we are not a matching service or a marketplace service. So because of our payer relationships, we have a substantially deep relationship with our therapists, both 1099s and W-2s, and we really treat them the same. And what I mean by that is we have a community of therapists. We have continuous medical education for the therapists. But most importantly, we hold them to a certain standard based on the five quality metrics, as I mentioned earlier, to provide therapy that is acceptable to our patient population. And because of our relationship with the payers, they audit us through that.
We have joint operating committee meetings with them as frequently as needed, if not once a year, if not more, where they actually look at the network and they look at our notes and they look at what's being submitted. So it's a really, really tough, very significant differentiator in the market. Of course, the other is we are really the only publicly traded mental health service that is in network with 200 million lives.
Yeah. No, thank you for that.
I would be remiss if I didn't say that the big strategic differentiator also for us is, which I think has been recognized, is not just our investment, but our capabilities on the AI side.
Yeah. And I was going to ask a question. Actually, I could go into that. I think one of the things that I took away from this was just the LLM that you're putting together and that's in beta mode, right? And so maybe we could talk a little bit into that. What do you think is the go-to market strategy for the LLM, Talk AI? And how do you just see broadly AI playing into your sort of broader growth into the next few years?
Yeah. So we haven't discussed or talked about what the impact will be relative to the financial part of the company. We're just not ready to talk about that yet. So in terms of the go-to market, there are multiple different, at least four or five different areas for commercialization. We will first go direct to consumer to test and learn who's using it, why they're using it, what does this pricing strategy look like on the consumer side. So there's going to be a lot of learnings that'll go once we go out to the market after the beta testing is done. So there'll be a lot of learnings there.
We do believe that, as I mentioned, relative to the populations that we think can use it, we think that there will be the possibility of significant opportunities relative to segments of the population, whether it's university or military or oncology or women's health, whatever it is, that those populations are definitely possible in terms of applicability. We don't know what'll happen on the employer-employee side because it will be a less expensive alternative to providing mental health services to employees. I think the licensing is a question discussion that's come up. Do other people want to use it? And to preempt the question, I think that at some point, the payers will probably get interested in what we're doing. Not right now, but sometime soon.
To you, even on the direct-to-consumer component, sort of who uses it and how and where they may migrate within the Talkspace platform, that's also something we're going to learn a lot from. For example, thanks to having been a pioneer in this space, the benefit of the brand awareness that we have of being in the market for over a dozen years, we have a lot of people who come to us, come to our site, actually start the registration flow, and for whatever reason, never end up sort of getting all the way, right? There's already a very, very large volume of people who come to us who are not yet ready to be face-to-face with an actual individual therapist.
Would they be more likely to engage with a, now with the proliferation of ChatGPT, something very familiar in app style, where whatever it is, the anonymity, the fear of judgment, some of those obstacles are not as challenging for somebody to sort of fully check out, if you will? For many of those people, that sort of what we're calling sort of lower acuity, subacuity care through the LLM will be plenty and absolutely satisfy sort of what they need, and they will stay there. At the same time, and this gets into the clinical oversight component, what percentage we don't yet know, but many we are sure will, it'll be appropriate for them to actually be referred into in-person care, right? So in some ways, it's sort of like a gateway into therapy and also going the other way, right?
Somebody may have been in therapy for a long time, and this LLM can act as sort of like a maintenance plan of, "Hey, I no longer need to see my therapist weekly or every other week, but just in case," make up your example, "my anxiety flares up again, I have this tool to turn to." The third option, sort of going back to the personalized podcast, where we've made this available to existing therapy members and saw a really nice retention uptake thanks to this engaging tool as an in-between session sort of way to engage with therapy in a form that's much more exciting than sort of your typical journaling exercise, if you will. There's also the potential that we would make this available to existing therapy members as sort of an ancillary tool to their in-person therapy.
Thank you, Ian. No, I appreciate that.
I think even just as general relational component, I know plenty of people who will talk in ChatGPT, and I'm like, "Why are you asking ChatGPT something about, go talk to someone?
It's very funny. A lot of thought leaders have been very upfront about this, like Eric Larsen, for example, where humans lie to other humans. They are not afraid to tell the truth to ChatGPT. That's been a really insightful sort of finding of this whole proliferation the last years. Even just in our market research about who might be most interested in this, it's very funny. The sort of generalizing demographic of that consumer proactively searching for AI therapists actually looks very, very different than our sort of typical consumer, which tells you, from my standpoint, this will likely be very TAM expansionary for us as opposed to an either/or kind of question.
Thank you. And I guess even just like double-clicking a little bit more here. So if you're in the app and you're getting your therapy session or you're just talking throughout the day, and it then does refer you and say, "Hey, maybe you should talk to a therapist or mental health provider," does it help you maybe fill out the forms or anything like that? Does it make it easier?
Exactly. We'll make it, and that's a key part of the product. Maybe John can speak to it, but that's, we'll make it very easy to funnel out.
Yeah.
Yeah. I remember this is positioned to have a conversation. It's not positioned for someone who wants AI therapy. I mean, they don't provide that, but it's more positioned to someone who wants to have a confidential, private conversation about something that's bothering them, knowing that their information is protected and that nobody's going to get to it. That's how it's positioned. Now, eventually, if he ends up in therapy, then yes, the event we think we know an advantage for Talkspace is if they do need therapy, we can very quickly determine eligibility. We're not going to turn them away. But by the way, the therapy that they need and the therapist is essentially free or covered service. So the other general-purpose LLMs, even if they were in some sense going to refer or figure out what it is, we're not sure how they're going to do that, right?
We are a covered service, so it makes it, to be honest, it makes it even stickier.
Exactly. Yeah. Thank you for that. I'll just look into the audience if there's any questions, particularly about the AI component of this, the innovation of the company. I think as I've listened to your presentations over the last couple of years here, I've always seen that you've always innovated, right? I think last year was the.
Podcast.
The podcast, exactly. Yeah. And so there's always just something new coming from the team that's like, how do we use technology to reach people in a way that works for them versus trying to get them to follow our model. We're trying to give them?
Well, the question, the history of Talkspace way before my time. Really, we're cutting edge in terms of making the recommendation and doing all the research that texting and messaging was adequate therapy. That was Talkspace, so the company is based in technology and innovation, right? We happen to be a healthcare provider of mental health services, but we do straddle being an innovative technology company on top of healthcare. We get asked all the time, are you a healthcare company that has innovation in IT or you're an IT innovative company that happens to be in healthcare, and the answer is we're both. Always happens, so.
Thank you. Moving on from AI, if there are no other questions, one of your primary KPIs, completed sessions, started to accelerate in the back half of the last year, 2025. Can you just maybe give us some sense of what do you think is underpinning that? What's kind of driving that growth?
It's such a big question. I mean, it's really multifaceted, but this is where we, as a management team, focus very data-driven weekly, right? So a lot of it's new user growth, which is coming from, as Jon alluded to, without ever needing to win another logo, we have a massive white space opportunity within our existing base of covered lives. And the largest challenge we have or sort of greatest detriments is to get the awareness out, which, again, fortunately, having had a history as a consumer-driven D2C business with deep marketing expertise, the brand awareness is very high. We have that slide, especially relative to what we spend. So we're able to sort of leverage this sort of asset in terms of the brand value we've created over a decade.
Where most people know of us, what they don't know is that it's actually a covered benefit now. So if you come across any of our marketing, which I'm sure if your phone's on airplane mode, maybe you will now, you'll see that the core message we lead with is, "Check your coverage, check your coverage." So no matter how you get to our site, we will push you down a sort of singular funnel, which is, "Hey, give us your insurance information because this is like," and we know that cost historically is the largest obstacle to getting care, which we've effectively solved, and it's just a matter of communicating that. So that's through marketing, and then as Jon mentioned, additional referral relationships, and then other sort of call it affiliate partnerships, like the Amazon partnership, the Zocdoc integrations.
I would also say, having been in many ways sort of a first mover in this payer strategy pivot, our view is that that first mover advantage is really crystallizing into a sustainable competitive advantage insofar as we've had a long history of working with the payers. Again, the differentiation that we are not a marketplace. We're not an MSO. We are a full-fledged HIPAA provider, right? All the care is happening on our platform within our sort of four digital walls, and being a true partner to the payers, they've really appreciated that. I think proof of that appreciation is coming through the fact that we're getting deeper and deeper embedded with the payers through these embedded directories where, as you can imagine, the traffic we see coming from the insurance portal route is very, very high intent and so very, very high converting.
And so working in partnership with the payers to make that sort of checkout experience as frictionless as possible has been a very big benefit. We talked about it on our last earnings call in October in 2025, and we have a number of new integrations to a similar effect in early 2026. So again, another way just to reach existing people who want to get to care are covered by their insurance plan to see Talkspace, but may just not know it.
Got it. Thank you. And I think it kind of almost segues into a question I had, and this was going to be more so about where do you see Talkspace both in the medium term and long term? But I'm kind of, as I hear you speak, I'm seeing it's going to be more integrations, more strategic partnerships, making sure that we can get customers to have frictionless conversions into the product itself to be able to access the resources and help they need without as little barriers as possible.
Correct. Yeah.
I guess, could you maybe speak to what that medium or longer-term vision is that you would see that as?
One, we will continue to improve the patient journey. I compare it to, "Look what it was, OpenTable 10 years ago to book an appointment," right? We need to make it as frictionless as possible for patients to find us, book an appointment, stay on the platform, and get better, right? That's our primary goal. It's always amazing to me. We go through every month, someone comes up with, or often some other different idea to actually improve the journey. You would think, "Okay, we've done it, we're finished." It's never finished. It's amazing how many more changes can occur to make it even better. That's a huge goal of ours, is just to, and that's a really big difference between us and a consumer. We are really, how do we get people, keep them, and keep them on the platform?
So I think that's where we're focused. I'd say on the rest, it just depends on the AI we're talking about. But so for instance, on the teen side, we continue to talk to lots and lots of other entities and school districts and counties and cities and states about how we can improve mental health services to teens. It's a really significant initiative for us relative to what's going on, so.
Thank you.
If I could just add on.
Sure.
Historically, we're known really as a therapy service provider. Over the course of 2025, whether it's through the Wisdo acquisition and sort of bringing on peer-to-peer and coaching, sort of going down the acuity scale, or sort of relaunching early in 2025 our psych business. At this point, we're much more of sort of a full-scale platform for a user and in improving both the member journey, but also the provider experience to be able to collaborate with other Talkspace providers, maybe in different areas, right? So psych to therapy or therapy to psych and making sure that it's a very warm handoff and notes are shared. Historically, that was not an easy process, right? So that's been a big initiative, which Jon alluded to the 50% growth we've been seeing in psych the last couple of years.
That's in spite of sort of that lack of communication internally across the platforms. Looking out to 2026 and 2027, there'll definitely be some revenue synergies, if you will, from that.
Got it. Thank you. Thank you. I know we have under two minutes left, so I'll open it up again for the audience if there's any questions. And if not, Jon, Ian, any final thoughts, words you want to leave us with?
I would reiterate that we've had a great three-year journey and a market that continues to be just enormous relative to the need. I would reiterate the fact that hundreds of millions of people have gone on to chat agents to get therapy. This is really an inflection point, quite honestly, for this industry in some respects. It's very, very different than it was a year ago, and I like to say we're in the, I know we're in the middle of what I'll call that vortex right now, right in the middle of it. It's unknown about how this is going to play out, but quite honestly, we've been successful before and we've purposely positioned ourselves to address an issue, which, as I said before, which in some sense to me is very positive.
I mean, now you have a lot of people who can have lots and lots of people who could gain access to an issue that's really dominant. So we'll see. Stay tuned.
Jon, thank you for your time.
Thank you very much.
Thank you, audience. Thank you for coming out and have a good rest of your day.
Thank you.