Healwell AI Inc. (AIDX) Earnings Call Transcript

Thank you for joining Healwell AI's 2026 Second Quarter Financial Results Conference Call. This call is being recorded. [Operator Instructions] I'll now turn the call over to Mr. Hefton Seni, Investor Relations at Healwell.

Hello, and thank you, operator. Joining me on the call today are James Lee, CEO of Healwell; Dr. Alexander Dobranowski, President of Healwell; and Anthony Lam, Healwell's CFO. I trust that everyone has received a copy of our financial results press release that was issued yesterday. Listeners are also encouraged to download a copy of our quarterly financial statements and management discussion and analysis that was filed on SEDAR+. Please note portions of today's call, other than historical performance, include statements of forward-looking information within the meaning of applicable securities laws. These statements are made under the safe harbor provisions of those laws. Please refer to yesterday's press release and to our management discussion analysis for more details on the company's risks and forward-looking statements. We provide forward-looking statements solely for the purpose of providing information about management's current expectations and plans relating to the future. We do not undertake or accept any obligation or undertaking to release publicly any updates or revisions to any forward-looking statements to reflect any change in our expectations or any change in events, conditions, assumptions or circumstances on which any such statement is based, except if required by law. We use terms such as gross margin and adjusted EBITDA on this conference call, which are non-IFRS and non-GAAP measures. For more information on how we define these terms, please refer to the definition set out in our management discussion and analysis. There will be a question-and-answer session at the end of the call, which will be limited to analysts only. [Operator Instructions] And with that, let me turn the call over to Healwell's CEO, James Lee.

Thank you, Hefton, and thank you, everyone, for joining us today. Before Anthony takes you through the numbers, I want to spend a few minutes talking about something equally fundamental. A year ago, we announced our first quarter post the pivotal Orion acquisition and embarked on a 2-year transition to integrate our businesses, embed our AI capability and shift our sales mix to more enterprise healthcare sales from episodic life science sales. We knew we had a great opportunity in front of us, and we're excited about the possibilities. But as I sit here today, we are feeling even more confident both of the opportunity, but more importantly about the progress we are making. We are through the difficult part of integration and transition, and we are now seeing the benefit of the enterprise healthcare focus. We have successfully demonstrated an upsell of customers with our AI capability. We've integrated our business lines, and we have improved our balance sheet. We expanded our AI footprint, and now we are moving to the exciting phase of delivering on those sales. At the same time, we've expanded our growth channels so that today, we have the strongest pipeline of our potential deals we have seen. Importantly, we have, at the same time, expanded our scientific validation mode, begun investing in infrastructure to improve margin and maintained a positive adjusted EBITDA throughout. The market we are building into is changing faster than any other point in the history of our company. And what we have built is we are building a business position to meet that need. I'm excited to share this progress today, but more importantly, excited to talk about what we will be demonstrating in the coming phase. Our mission is to be the primary enabler of preventative care. But what does that mean? So we're going to talk about it in 3 simple sentences. Firstly, we connect. We take complex fragmented healthcare data and turn it into longitudinal patient records that can be used to share across the system. We surface. 80% of clinical data is unstructured, sitting in notes and reports that were never designed to be read at population scale. We abstract the clinically relevant information from that data with 95% accuracy and 95% of disease states, and we enable. We give the healthcare system the ability to implement preventative care cost effectively, not as a pilot, but as an operating model. Now the facts are really well known, so I won't go into them, but I'll give you more about the shape of the number themselves. So 3 of the world's largest healthcare systems all face 3 very different pressures, but they've all faced with the same outcome, which is the money is effectively gone. Whether it's the debt servicing overtaking healthcare spend or the working age population being too small to carry the load, every one of the major healthcare systems around the world are running out of room within this decade. Underneath that sits a disease burden that no one's counting because most of the people who have these diseases don't know they have it. And here is the point that matters commercially. Each one of these diseases is detectable many years early before the costs are rising. The signal exists that is simply not being read. 4 forces are converging, and converging is an important word. Any of these on their own would just be a trend, but all 4 occurring is a structural shift. The economics are broken. We've talked about that, and that's a well-known fact. So systems are having to intervene even earlier to ensure that they can spend less. The regulation is starting to catch up. Prevention is now mandated, not just aspirational with data access and interoperability rules that are a direct run rate for our platform, and the market is aligned. The payers, providers and pharma are all reaching the same conclusion. The opportunity is that these diseases are detectable early, prevention costs a fraction of the treatment and our platform today connects the data to science into delivery. You will hear us talk a lot today about the transition from episodic to enterprise because that is the part that we've been focusing on the part that matters. Whereas last year, we were negotiating 6-figure individual engagements. Today, we're discussing annualized licenses that exceed that. The old model was project by project services and patient reviews. Revenue grew in a linear fashion, one clinic, one study at a time. Delivery was high touch and manual. Revenue was lumpy, nonrecurring and hard to forecast. Our current model we're moving towards is different in every single way. We've go to enterprise licenses, data access and a productized SMARTSuite. We deploy it once and expand it across our network. We're selling to global pharma centers of excellences. Delivery is now product-led, revenue is larger, stickier and multiyear recurring and the defensibility of our data activation plus the DARWEN platform is a real moat. And importantly, we're now through the trough of that transition. Now that transition has produced 4 key growth engines, all from the same platform across 2 key client segments. Firstly, engine 1, our HIE and clinical data unlock. Unlocking de-identified data across our partner networks creates value for every stakeholder right across the ecosystem. We are seeing active interest from all markets in this capability. Engine 2 is our SMARTSuite for its Search, Summary, or ID. It's packaged, it's repeatable and it's embedded into our platform. It's a single sale process and it's been successfully deployed in North America, sold in the Middle East, and we look forward to giving further progresses in the second half. Engine 3 is our global life science enterprise business. We're moving to centralized agreements with pharma AI centers of excellence, which is a cost saving for them against repeating patient ID and real-world evidence studies region by region. And Engine 4 is our consented data in Canada. We've talked a lot about WELLTRUST, but activating this key partner network for life sciences, CROs and public health has created a growing consented database with strong demand behind it, which has become a virtuous cycle. The more consents we get, the faster our customers can innovate. So a year into this transition, the execution signals are good. They're shown across all of our geographies, products and partners. In '25, we were predominantly a life science business with episodic revenue. Today, we have activity in all of our key markets with healthcare systems, and this will grow into a meaningful part of our revenue. Excitingly, the integrated offering of DARWEN and Amadeus is commercially compelling, is resonating with all of our customers. WELLTRUST is opening genuine new commercial use cases and moving us towards meaningful enterprise discussions across new revenue streams. SMARTSuite now means we sell as a single customer and a single company through one sales process. With an AI landing expands the customer base these conversations have transitioned from project work to enterprise deals and the platforms and businesses are working well behind that. Finally, data activation. This brings us closer to the customer and reduce duplication because we are helping our customers activate data they already hold. There's no new data assets to build, which shortens the pathway to evidence. Internally, AI data output overtook human output for the first time in May. From a base of 0 a year ago. So we're seeing significant AI use case internally. Commercially, we've had wins across the Middle East, Canada and the U.S., and we're seeing demand from all of our key geographies. These sales that we're winning are durable. They take a long time, more than a single quarter, but we're moving to an enterprise license with SaaS fees, and this is a shift that matters. I want to touch briefly on why the validation compounds across segments. So I think that's underappreciated across the market. Obviously, we're focused on science first with 57 peer-reviewed publication and pre-clinical recognition. But what this means is the healthcare systems can adapt preventative healthcare using validated evidence rather than assertion. And every adoption generates more data, which feeds into Amadeus and adds to 150 million lives. That scales what comes next is that allows something that underwriters can price because we can price that and we can move forward to the underwriter population. This is why we see insurance as the next and largest step in our opportunity, and we'll be talking to that in the year to come. Finally, I want to talk about our 4 key strategic goals for the quarter. As you know, our indirect stake in SpaceX is now estimated at approximately CAD 23 million as at the 30th of June 2026, up from a carrying value of approximately $4.6 million as at March 31, 2026. Our holding remains subject to the customer pre-IPO lockup period. We continue to commercialize our AI product suite across our care network. Our AI solutions include SMART Search and SMART Summary and they've been cross-sold into the Orion network now. We completed a real-world evidence study demonstrating the effectiveness of WELL AI Decision Support, identifying patients who may have been undiagnosed or unmanaged diabetes. WAIDS analyzed patient records and recommended clinical reviews and point-of-care assessments for high-risk patients. We completed a multi-province pilot evaluating our DARWEN SMART Summary and Search solutions across the healthcare systems across British Columbia, Ontario and New Brunswick across both OSCAR Pro and Intrahealth Profile EMR. The results of that pilot were accepted for presentation at the American Medical Informatics Association, which will be taking place in November 2026 in Dallas, Texas. I'd like to hand over the call now to Anthony Lam to walk through our Q2 numbers.

Thank you, James. Before I begin, I would like to remind everyone that all of the figures I will be discussing on our call today are expressed in Canadian dollars, and our financial statements are presented in accordance with IFRS International Financial Reporting Standards. Our second quarter 2026 results as compared to Q2 of 2025 are as follows: Healwell achieved quarterly revenue from continuing operations of $33 million in '26 -- in Q2 2026, in line with the $33.2 million generated in Q2 of last year. The comparison reflects an unusually strong Q2 '25, which benefited from elevated project-based professional services activity, while our recurring subscription support and maintenance revenue continued to grow year-over-year, reflecting our continued transition from episodic project-based engagements to higher-margin recurring enterprise sales. Healwell reported positive adjusted EBITDA of $1.1 million in Q2 2026 compared to adjusted EBITDA of $2.3 million in Q2 2025. Our prior year performance includes Mutuo, which was disposed of in Q4 of 2025, together with a stronger number of life sciences studies in Q2 2025 accounted -- this accounted for the change in year-over-year EBITDA. Healwell achieved gross profit of $17.9 million during Q2 2026 compared to $18.7 million in Q2 2025. The decrease is due to a lower margin mix of studies within the Data Science and AI segment in the current year period. Healwell achieved a gross margin of 54% during Q2 2026 compared to 56% in Q2 2025. Looking at our first half of 2026, Healwell achieved revenue from continuing operations of $66.2 million for the 6 months ended June 30, 2026, an increase of 60% compared to $41.2 million generated in the 6 months ended in June 30, 2025. The increase was driven primarily by the Orion Health acquisition, which closed on April 1, 2025, and contributed a full 6 months of revenue in the current period versus a partial period in the prior year. Healwell reported positive adjusted EBITDA of $1.9 million for the 6 months ended June 30, 2026, compared to $0.1 million in the prior year period, an increase of approximately 2,574%. The increase was primarily attributed to the Orion Health acquisition and improved performance across our operating segments. Healwell reported positive EBITDA of $1.9 million -- sorry, Healwell achieved gross profit of $37.4 million during the 6 months ended June 30, 2026, an increase of 62% compared to $23.2 million in the prior year period due to higher revenues driven primarily by the Orion Health acquisition. Healwell achieved a gross margin of 57% for the 6 months ended June 30, 2026, compared to 56% in the prior year period. A key highlight this quarter is the continued positive trajectory of our operating cash flow. We generated $4.5 million of positive operating cash flow for the first half of 2026, a $14.4 million improvement, approximately 145% compared to cash used in the prior year period, reflecting an underlying strength in increasing efficiency in our combined operations following the Orion Health acquisition. This shift to sustained positive cash generation strengthens our balance sheet and liquidity position, and we remain in compliance with all of our covenants under our credit facilities as at June 30, 2026. We continue to prioritize disciplined capital allocation as we build on this cash flow momentum and work towards sustained profitability. With that, I'd like to now turn the call over to our President, Dr. Alexander Dobranowski.

Thank you, Anthony. Thank you, James. I'd like to take a moment to walk through the scale of impact of Healwell's platforms as we deliver globally. This is really the proof point behind everything James and Anthony have just walked you through, and it's one thing to talk about an integrated platform and another to show you the reach it's actually achieving across patients, clinicians and the broader healthcare system. Let's start with data. Across our global business units spanning 13 distinct clinical data domains, Healwell now maintains more than 150 million patient records. That breadth of longitudinal data is really the foundation of everything we do. It's what allows our AI models to identify at-risk patients earlier and more accurately than a single source system ever could. On the clinical side, more than 770,000 clinicians and physicians are now engaged across our platforms globally. That's a meaningful and growing share of the frontline healthcare workforce actively using Healwell software and technology in their day-to-day practice. And that engagement is translating directly into outcomes. In the second quarter alone, our AI copilots powered by our DARWEN AI engine identified over 62,000 high-risk patients. Patients who, in many cases, have not otherwise been flagged for early intervention. On the clinical technology side, our DARWEN AI engine now supports and is able to screen for 123 rare complex, and chronic diseases, reflecting the continued expansion of our clinical validation work and the depth of our disease detection capabilities. From a network perspective, we serve more than 22,000 healthcare service providers across our combined network of platforms and business units, underscoring just how embedded our technology has become across the broader healthcare ecosystem. And finally, on adoption, as of the second quarter, 1,291 physicians have been onboarded to our AI copilots, giving us a strong and growing base of active physician users to build on as we scale. Together, these figures reflect the scale and reach of the platform we've built and more importantly, the depth of real-world impact we're delivering for patients and providers around the world. This is the kind of scale that we believe differentiates Healwell, and it's a big part of why we remain so confident in the growth opportunity ahead. From an outlook perspective, we remain focused on several key drivers of growth as we continue to build on the momentum we've established this quarter. First, on profitability. Management and the team, we continue to target an approximately 10% adjusted EBITDA margin by the end of the year, reflecting continued operating leverage across the platform as we scale our revenue base and drive further efficiencies across our combined operations. Second, on revenue quality. We expect, as James reflected, a growing mix of enterprise recurring revenue led by subscription, support and maintenance growth, which continues to make our overall revenue base more durable and more predictable quarter-to-quarter. Third, on deal economics, we now anticipate larger AI deal sizes as SMART Search and SMART Summary scale across our customer base, reflecting both deeper adoption within existing accounts and the growing maturity of our AI product suite. Fourth, on geographic expansion, we continue to see organic growth in new customer wins across our key geographies, and we expect that footprint to keep expanding as our enterprise pipeline matures. Fifth and finally, on the public sector, we're seeing extremely strong tailwinds supporting continued adoption of our platform as governments and health systems increasingly prioritize preventative data-driven care. Beyond these 5 priorities, we also continue to see meaningful key value and growth alongside WELL Health, our strategic partner across 3 specific areas of that relationship that James highlighted earlier. First is WELL ID, our secure digital identity and single sign-on platform across the WELL Health network, which continues to streamline clinician access while strengthening the security and interoperability of our broader platform. The second is WAIDS, WELL AI Decision Support, our AI-powered clinician decision support platform for early disease detection, which continues to generate strong clinical validation and commercial momentum. And third is WELLTRUST, our patient consent and data governance infrastructure that underpins secure data access across our platform. Across all 3 of these areas, we're looking to expand these capabilities into additional geographies in the periods ahead, and we see meaningful runway to deepen this partnership even further. So taken together, the profitability discipline, the shift towards recurring enterprise revenue, the scaling of our AI products, our geographic expansion and public sector tailwinds and the depth of our partnership with WELL Health, these priorities reflect our continued discipline in scaling a connected AI-enabled platform. It's a platform that we believe is built to deliver durable long-term value for our shareholders, for the patients we serve and for our healthcare partners around the world. In short, 2026 has been a year of focus on margin expansion and improvement with 2027 being a year where we will focus again firmly on growth. And with that, I'll now hand it back to the operator and move to the Q&A portion. Thank you.

[Operator Instructions] Your first question comes from Kevin Krishnaratne from Scotiabank.

Congrats on the continued success. I wanted to talk about your mention there of moving from episodic to enterprise. Maybe it's in public health, actually. Wondering if you talk about the current pipeline of opportunities, what's that looking like on potential HIE RFPs? Maybe you can talk about opportunities by geo. And just curious how the discussions with customers are progressing? Is there any change in the macro or healthcare budgetary front that we need to be aware of as we think about the timing of potential RFPs and how that would translate to revenue?

Kevin, look, great question. You probably asked enough in that question. I could probably spend the next 30 minutes covering off. So let me try to bring that back and then just maybe ask a follow-up and push me if I could get some of it wrong. But like starting your first point, it's not just the software, it's just not in healthcare. We've actually shifted all of our sales to -- sorry, all of our focus to enterprise sales. So we're seeing potential deals not with our life science partners, we have data unlock partners across HIEs and across our SMARTSuite Suite with enterprise sales. The pipeline now would be that we're seeing opportunities in all 4 areas. But importantly, we're seeing it across all geographies. And so it's very hard to sort of isolate it down. What we've actually seen in the last 6 months is a shift in approach from the market, partly because there's been so much -- so many people trying to do small pilots that what they're really looking for is much more enterprise platform outcomes, so not a single point solution. So what we're actually seeing is that the conversations are morphing much more quickly from a little pilot to show us what a proper RFP would look like at scale. To give you some context of size, a pilot, we've talked about previously might have been $200,000 or $300,000. What we're seeing now is that an enterprise agreement might be anywhere between $1 million and $3 million, depending on the size of the customer. And as I said, it's across all parts of our business. Geography-wise, we've seen a lot of activity in the Middle East, a lot of activity in the U.S. Canada is busy and Europe, we've got some RFPs coming out towards the end of the year. And importantly, for home markets, both Australia and New Zealand are busy at the moment as well. So it's quite broad-brushed. What we've seen is -- it's across all of our product suites and across all of our geographies currently, which is why probably some of the optimism we're seeing currently in our language. Did I leave any out there, Kevin?

Yes. No, I think you captured it well. And maybe to get Anthony on board here, just as we think about the model, thanks again, good to see the confidence in the 10% EBITDA margin trajectory. How do we think about your thoughts or any guidance you want to give us on revenue growth by segment on the Data Science and AI and then the healthcare software for 2026? And maybe it's a bit early, but if you can just talk about the trajectory into 2027 on your growth expectations?

Yes, Kevin, great question. As we look out here and as we talked about, the shift to our -- to more enterprise sales is going to create an elongation of sales cycle for us. And so while we started the year with a very exuberant sense that we view that 30% to 50% in terms of growth, we're likely to be more on the lower end of that growth range in terms of 2026, but really see 2027 on the heels of the health system side of the AI and DS part of the business to be a big contributor of our growth in 2027. So Kevin, to your question on AI and DS, I think we can see that growth on that profile has probably shifted to the right for us from our perspective, given our real focus on enterprise. And then as we look at healthcare software, I think growth rates that we've been talking about to now, I think you can expect to see continue in that high single digits kind of range is really as we look at that business as a steady state kind of really stable core revenue stream.

Your next question comes from Gianluca Tucci from Haywood Securities.

I guess, firstly, at a high level, can you speak to cross-sell attach rates so far for selling into the Orion legacy customer base? How is that tracking?

Yes. So our goal was to get 10% done this year. I think as we sit there now, we have line of sight as halfway through the year for that. So feeling really comfortable with our current target that we'll achieve our cross-sell. We've got another big push coming second half to a bit wider. But realistically, we're probably towards capacity of what we want to get done in the first year to make sure we do it well. Obviously, the enterprise sales cycle is probably more like 6 to 9 months, not 1 to 3. So we're seeing that way towards the second half.

Great. And then just perhaps a follow-up on the adjusted EBITDA margin target of 10%. That would imply a strong second half. Can you bridge that for us what kind of cost actions or like revenue combination helps to get you there by the end of the year?

Yes, there's 3 things you see there. Obviously, one is the enterprise sales starting to kick in, in healthcare and AI and DS. So we'll start seeing that occur in Q3 and Q4, given that they were small numbers in the first half that will shift that division materially in terms of margin profile. And cost actions we've seen taken in the first half flowing through in second half in the software business. And then finally, some continued growth in our software business items. There's obviously some timing issues between first half and second half within costs. So first half costs might have been slightly higher than second half. But effectively, those 3 things bridge that gap.

Next question comes from Michael Freeman from Raymond James.

Congrats on the quarter and the progress. I wonder if we could double-click on your pursuit of the insurance opportunity. I wonder if you could frame that again and sort of update your view on it from when you introduced it, the pursuit of that last quarter?

Yes, sure. I guess when you look at the framing of it, the best way to think about insurance is around -- in the U.S., obviously, insurance replaces where the role of public health does in commonwealth countries, i.e., the fundamental payer. One of the things we're finding in the U.S. is that the insurance market has a secondary use of risk, i.e., so looking for where there are gaps in care where they might get sued. And so what we've seen in the U.S. is there's 2 different lenses. It's not only cost savings from an insurance point of view, but it's also risk mitigation. And so it's got a really different driver in that market. And what we're also seeing in the U.S. in particular, with the insurance side is that they are tied to the provider network. So we're also seeing that where we're talking to providers in the U.S., they've actually got connectivity to the payer markets. And so we see those sales as much more intertwined. What I would say is they are larger but slower burn sales. So we won't be expecting to make any of those in 2026. But the sheer size of that market is what we're excited about, Michael.

Okay. All right. Great. I appreciate you framing that. Now I wonder if you could touch on the balance sheet and how you're feeling about its profile and how you can anticipate it evolving in the second half of the year?

Yes. I'll open that then maybe Anthony, you can close it out. But I think the best way we think about our balance sheet, Michael, is that in terms of our short-term liabilities post the disposal of SpaceX, we'll see net liabilities in the short term of circa 1x adjusted EBITDA. Our long-term debt, $30-odd million, is a convertible instrument, as you know, with 4 more years to go. So we look at the balance sheet now as being a really stable part of the business to serve the needs we have today. But Anthony, is there anything you want to add to that?

Yes. Look, great question on that because I think the big thing that James highlighted on the call was that we -- with the -- our investment in SpaceX, I think, again, the intention is clearly to liquidate that position. We will actually have a good source of capital there for our immediate needs. And so the balance sheet actually for us is we feel very good about because while we approached cash flow neutrality and generation towards the end of the year, we're now in a very solid position from a liquidity standpoint with that asset being one that we will be liquidating to. We feel very comfortable that we have all the resources we need to meet the expectations that we've been setting for ourselves, not only for this year but also for next year.

Your next question comes from Brian Kinstlinger from Alliance Global Partners.

I appreciate your changing approach to get away from episodic demand and the opportunity as it relates to preventative care is clearly large. The subscriptions for AI technology are almost half what they were 2 quarters ago. And this is the part I'm focused on, not the services piece. So first, what's driving this reduction? Next, what's the biggest impediment you're seeing right now to growth? And lastly, what gives you the confidence the ramp is imminent as we've been talking about it for a bit?

Anthony, do you want to start on the first one? Because obviously, when I look at -- when I look at our services subscription revenue and AI and DS, I'm not seeing it down. So do you want to maybe comment what number you're talking about?

Sure. We did $531,000 in the June quarter for subscriptions. And just 2 quarters ago, you were at about $1 million, and it's come down each of the last 2 quarters.

Anthony, again, I'm looking at a very different number. I see our number is $479 in the MD&A. So...

Brian, if I could just point out, you're looking year-over-year. In our prior period numbers, if you're looking at the prior period numbers up until the beginning of the first quarter last year, we had a business called Mutuo that was -- we have divested. Those numbers stay in our comparatives because we were sold. But they were 100% subscription, and that's probably what's throwing off some of your numbers there. I would say, otherwise, our subscription numbers have been pretty steady for every other part of our business. And so I think that might be the skew for you in terms of what you're seeing in terms of the subscription piece.

Okay. Let me ask differently. Last quarter, when that business was not in the numbers, you did about $828,000. So we're down 36% sequentially. So it speaks to the same trend. Again, I'm curious, what's the biggest impediment growth? What's leading to the churn? And what gives you confidence that imminently we'll be growing this?

So Brian, I don't -- look, I'm happy to take this up with you. We didn't have any churn in the quarter. So the changes that we've had in our -- in any of our business is really around our episodic revenue or maybe a little bit on our professional services, but our subscription business remains pretty robust.

We can take that piece back up online. But in terms of what's giving us confidence in the second half is revenue recognition is effectively the key issue there. And so when you're delivering on enterprise licenses, the revenue recognition is very different from the sales that we've made. So we've announced sales in the Middle East, in Canada and the U.S., and we're in the process of delivering those. So we'll be able to recognize revenue against those as those delivery milestones are made in the third and fourth quarter. Does that...

Sure, I mean, hopefully, you can take it offline. That's good, but the numbers are down. But the -- in a previous question Anthony responded to, he said you'll probably be at the low end of the 30% to 50% growth for the AI segment. First of all, what number does that suggest for 2025 AI Data Science because what was reported was $10 million. And I'm sure that doesn't include the divested piece. So I'm just kind of curious what that suggests for the second half of the year?

Yes, Brian, we did start with -- so the comparable for 2025 is that $10 million mark. And as I mentioned, we're going to be on the lower end of the growth scale on that in terms of year-over-year growth for '26. And so that would suggest that we are in closer to that $13 million for the full year.

So that would suggest almost $9 million of second half revenue for AI and Data Science? From $4 million change in the first half of the year?

Sorry...

From a sales cycle, I just -- I mean, I guess maybe you can talk about some bookings that get you there already.

So Brian, in our first half...

You did $4.6 million.