Nordhealth AS (NORDH) Earnings Call Transcript & Summary

Hello, everyone, and welcome to our Nordhealth Capital Markets Day 2026. My name is Alex Cram. I'm the CFO and I want to start by just saying a big thank you to all of our in-person attendees for traveling to London to come and see us today. We have investors from all of the Nordic countries, the U.K., Germany, France, and so I want to say a big, big thank you to everyone who's traveled especially our guest speakers who've come from the U.K., but we really appreciate that they'll do a fireside panel for us later. We have a fantastic agenda for you today. We're going to start off by going into our group vision and long-term objectives. We're going to have a deep dive on the veterinary business and strategy. We're going to follow that up by a look at our AI products and our vision for AI. We'll take a short break at that point. And then after that, we'll have the fireside channel with two of the largest veterinary groups in the U.K. After that, we're going to go into our therapy business, and then we'll finish up with financials and guidance. Today's presenters are Charles MacBain, our Group CEO; James Stanier, the CTO of our veterinary business, Karan Wallia, the CEO of our therapy business and then me. And with that, I will turn over to Charles for our group vision.

Hi, everyone. So there are many new investors and there's current investors, which will be watching this. So I wanted to start with going through a bit of the basics of what we do, and then we'll go into more detail. There's a strategy. Let's start with what problem we're trying to solve. So -- the main enemy we're trying to solve is admin, right? Both in veterinary and in therapy, the admin burden has increased mostly due to regulation in both these specialties. And most of this admin work is unpaid. So our mission is to be able to reduce and automate as much of the admin as possible. And the mission is to give health care teams more time to care, right? More time to care for their patients or pet parents of NVS veterinary, more time to care for the business that they're trying to grow and more time to care for themselves as well. And that's what we're trying to do by building our software. In terms of the verticals -- we operate in veterinary where we do veterinary practice management software. We've got just over 4,000 clinics across the world and also in therapy, particularly psychology and physiotherapy, where we've got 13,320 clinics. And so we automate admin for over 17,000 practices every day. And we've got over 75,000 active users every day across our products. So I want to start from where we're positioned. So especially now that there's a huge amount of disruption happening in the technology world. So the practice management software, which is the product that we sell is the core of the practice right? We -- the majority of the data of the practice is stored in the practice management software. We own the data schema, right? We also are the place where the majority of workflows happened. So the people interacting every day with this data to try to input data and extract data or doing it via our services, right? We also have all the compliance feature built in for both veterinarian therapy, for example, in the permissioning. So if a reception has access to this data, but not this data and so on. We also have the a lot of trust that we're very thankful for from our customers when we build very closely together of them. You'll hear from 2 of them today on the veterinary side. But that's 1 of our key assets that we can work together with them to build great products. Next one is that we actually -- when I started the business, we were around EUR 3-and-a-bit million. And so we didn't really have the scale to be able to invest. Now we're north of EUR 50 million of recurring revenue. So we've got the scale and P&L to be able to invest in building a world-class products. And we also have lots of different practices where we can actually push things out to them, so to understand how they're working better. So we've got -- and the last thing is the team. James will talk a little bit about it. But we have a fully AI native team on the R&D side and even beyond, we see like our team members across other functions beyond R&D using AI quite a bit. And we've done a big restructuring of our product in both in veterinary therapy to make sure that we have a great team for this era. So -- it's always good to talk about here's the strengths. Here's what we can do, but we've actually done something about it. So we've launched a great AI Scribe for Veterinary and Therapy. We were also the first to market globally on the veterinary to launch AI billing, which is basically when you speak, it actually adds the items as well. We were the first in veterinary to launch an MCP, which basically allows you to query your data via cloud or OpenAI. And also, we have an ability to do this within the product as well as Provet. James will talk about this later. We've also been able both veterinary and also seen therapy gather all the data from all the different sources, there's PMS that was imported from the previous clinics. All this data is hard to be able to access when you have a 15-minute consultation, so we've been able to extract all the relevant information to gather a concise patient history. Another great thing is that one of the big issues that we've heard from vets is that the -- and pet parents is that they don't get the information they need. And so our AI discharge note enables to distill what's happened, the medical terminology into pet parents' language. And on the therapy side, there's a lot of forms, especially as you go in many different Nordic countries where there's a lot of regulation. And so a lot of the therapist time was spent filling in those forms that same information again, again, and they've launched this their form fillers. And where we're going next, James will talk a little bit about this. But for now, we're improving and automating the recording of information. So instead of people in practice being typists. They can just do their work and speak and most of that will be entered into the database. The second is summarization of information, right? The next step of it will be us proposing the next step based on the information that we have and then taking action. Obviously, with clinician control in many areas and being able to coordinate better across the whole work flow and also enable agents to coordinate within those workflows. The goal is the same, right? This is just an enabling technology, but our mission and goal is just getting time back for the clinician, getting time back for the receptionist. A little bit about Nordhealth at a glance. We reached EUR 50 million of signed ARR. We've been growing north of 20% over the last 5 years. Our EBITDA minus CapEx margin is still negative 5.3%. We've been investing a lot in AI and also in expanding into new markets. it's actually quite profitable to recruit new customers. Our lifetime value to customer acquisition cost ratio is around 16%, right? That's because we invest a lot in products because we believe product is the way to win, right? Our churn rate. This is the unique figure about our business. It's the best thing and also the worst thing. It's like churn rate of 3% means that people stay for us for 30 years, right, which is a crazy number, right? Because -- but it's also hard to recruit people from -- even though they're legacy software, it's quite hard to recruit those people to move over to you. And then the net retention rate. This is the amount that a current customer spends more every year of us. That's driven by two things. One is groups are buying other clinics, and they migrate them over, right? Secondly, we build new innovative products that we upsell to our customer base. So just a bit about the history. This where I bought the company in November 2018. Now we've been growing over 44% over this period of time. What I particularly like is the organic part. We -- 75% of this growth has been organic. It's been conquering one country after the next. Next, adjusted EBITDA. In 2021, we IPO-ed, and we decided to get much more aggressive with our expansion. So we invest a lot in the product and go to market to be able to expand in veterinary, in the U.K., U.S., Spain, Germany, and you can see that this normalized. In 2025, we are aiming to be profitable, but then we decided to double down and invest on AI. In that there was a big disruption that could have happened, and we see this as a huge opportunity to be able to accelerate our mission of giving time back to care. There's four main things we're trying to do, right, and across two buckets. One is our customer mission, which is making sure that practitioners save time. right? That's our #1 mission. The second is that how is by making the software do the work versus people manually doing the work. The where is, we want to go -- we're going into the U.K., Germany and U.S. on the veterinary side. On the financial mission, right, we want to grow a discipline, we want to make sure that we're profitable starting in 2027 and every year thereafter. Second is 15% recurring revenue growth, plus or minus 2% as a group over the next 3 years. So that's a bit about the overall I'll do a little bit of a deep dive on the veterinary business next. Where is Provet today? So I'll go back to the same framework that I had for practice management software. On the data side, we've got 4,320 clinics using our software every day, right? On the workflow side, there's over 21,000 vets using our software every day, right? And on the compliance side, we've got very, very -- probably two specific permissions for role for markets so that the different corporates can decide, do I want to be fully centralized? Or do I want to be fully local for local economy? And you've got that ability to do that in Provet the fourth is trust. I think the best proxy for trust, I'd like to use is churn. And actually, the veterinary business has even lower churn than average at 2.2%. So -- and then the scale. We actually treat 1.7 million pets every month from Provet. R&D, James will tell you more about it, but we've been investing aggressively in building up these AI tools because we see massive potential to be able to meet our mission. We've got six agents live and looking forward to iterate on those and build new ones. In terms of the market in front of us, there's around 53,000 clinics in our target markets. We've got 50,000 left. That's a huge amount of clinics. In terms of TAM, right? We've got EUR 29.7 million, and our total TAM is around EUR 570 million. So we're still a very small player just in our target markets. What we've seen is that in the Nordics, our EBITDA is CapEx margins can be around 40% at scale, right? We are aggressing very aggressively in our new markets, but that is not in the margin at scale. A bit more about the market. We have around 58% of small animal clinics in the Nordics using our software. There are some competitors, all very local, mostly less than 20 people in those companies. In the growth markets, we only have a 4% market share. You can see in the U.K., we've got around 14%, 4% in Germany, most of which are on Vetera, the software we acquired, 5% in Spain and 1% in the U.S. These are the focus markets for us going forward for the next 3 years. I want to break down a little bit more about where the growth comes from, right? In terms of market opportunity. So the total TAM is EUR 570 million, which is 19x where we are today, right? We're half of -- more than half bit is from the U.S. But we've got a significant TAM as well in -- left in the U.K. and Germany and Spain. Where we -- in the U.K., our strategy is to go after the corporate was here from two of them today, and there's four others as well in the U.K., large ones and same strategy in Germany, where the corporatization is much lower or probably around 10% versus 60% in the U.K. Starting that journey with them. The U.S. as well is also a corporate first strategy, where we are -- there's around 40 different clinics chains there that are not yet consolidated on one software. In terms of unit economics, right, last 12 months ending June, we grew 13.6%, right. Our net retention rate is 109%. This is mostly about clinic chains rolling out net new locations. Churn was 2.2%, as I mentioned. Our LTV CAC is 26.7 which is quite healthy. The amount we spent on sales and marketing relative to is quite low versus the revenue potential. As we mentioned, the EBITDA-minus CapEx margin around 40% Nordics. We're now the #1 player by far in Europe, probably doubled the next and #2 in the world after ezyVet, which is owned by IDEXX. So in terms of the revenue in 2020 -- well, 2018 when I started, we're mostly a Finnish business. We used to have English classes, I remember on Wednesdays in the office, right? So it's a very, very Finnish business, then we conquered the Nordics, one country after the next, probably too many countries at the same time. I've learned that hard away. But -- and in 2022, we both the majority of the Nordics using Provets, right? And we expanded internationally in the U.K., Spain and U.S. Now over 51% of our revenue comes from beyond the Nordics. I still remember showing this chart where we had just Finland, and the Nordics was the expansion. That's -- so we're looking forward to not adding new countries and focusing on those to make sure that we provide a great product for those countries. Then the second part is our enterprise growth, right? So the Nordics was one of the first parts of the world to be corporatized. A lot of the big corporates actually had their first roll-ups be in the Nordics. We benefit from that and that those were our early customers. We really built Provet to be an enterprise first product. You can see this translation where over half of our revenue comes from -- PMS revenue comes from enterprise customers. We do also have additional revenue of EUR 3.9 million from payments and other partners, which is a new business we've built. But the driving force, and we see going forward will also be our enterprise partners. In terms of growth, we have invested more in 2025, but we still remain both EBITDA-minus CapEx and EBITDA breakeven at least. And the slowdown in our rate of profitability in 2025 was a deliberate choice, right? Because we don't want to be profitable, but not have the best product in the market. We are committed to when there is an opportunity for us to accelerate our mission, we will go after that opportunity to make sure that our customers get the latest and greatest. So -- how do we differ from our different competitors? So compared to incumbent PMSs, right? One is we're independent. When choosing us, like we are committed to building the best software that can save you the most time. We don't care if we sell you more lab test or we sell you more -- get more volume for you on your wholesalers. We are committed just on that one mission, right? Second is we have made a big pivot to become an AI company. James will talk a bit more about that. The third is like we don't talk about having agents, we actually built them, right? Then you can judge us based on the number of people that use us, those new features right. There are some start-up PMSs in the U.S. and the U.K., for example, and the big differential of us versus them is that practice management is hard. The best thing was the churn, the worst thing is that churn. If you're a VC-backed company, and people don't churn, and it's very, very hard to be able to gain that scale. The financial stability is not there. So at some point, right, they might be successful, but there's a big chance that there the amount that they spend, they lose every year. If they don't get the next round of funding, they might go away. Then relative to the start-up PMSs, we spend a lot of time on building secure products, right? The threat of security is increasing every day, right? The next is that we've built an enterprise-first product with centralized reporting, centralized pricing and item management. We are proving at scale. We have actually run hundreds of clinics on the same database. Relative -- then there's new competitors which are AI scribed, their point solutions just doing the clinical workflow. Yes, they do have -- they are AI native. They've got good agents. But you have to switch between one after the next, right? They're only point solution. They're not only one solution. They also could go away because they don't have the financial stability and they don't have all the different enterprise features required. That's how Provet is positioned in the market. Now the two growth levers, we say, win new locations, right? So we've got 3,600 locations on Provet, 5,360 -- and the second is, as we add net new products, we can grow revenue per location from around 570 to 1,900. I'll go into each of these separately. So -- the first is on locations. Number 1 is the U.K. We've got around EUR 6 million of ARR today, right, which is around -- and the total TAM is around EUR 60 million, right? We've got 14% location. TAM is higher because of the AI adoption rate as well, increase the TAM. As you can see, 16% of the market is owned by the top 6 players and that's what we're mostly targeting in order to win the U.K. Second is Germany. We bought Vetera in January 2022. We left it there because we wanted to focus on the U.K. and implementing our corporate partners there. We've got around EUR 2.4 million of ARR today on Vetera. The total market is around EUR 80 million, very little -- a few clinics are actually owned by the Big 5, they're around 4%, but it's growing quite fast. This is a market that's very underserved. The majority of the clinics are on server-based software. They actually have a server in the clinic, and it's the 21st century. Germany is normally quite a laggard on the side, but over time, they will shift over their cloud provider. There's no current cloud provider in the market. Then the U.S. So focus on U.S. enterprise. It just repeated the playbook that we've had in Europe that's been successful for us. The nice thing about the U.S. is -- and what I want to -- yes, it's a really tough market, but the expectations are high. If we do this, it also benefits all the rest of our customers in the rest of the world because they're fussier right? The at least, I would say like fill-ins the happiest place in the world, but it also has the lowest expectations, right? And so that's why there's a happiness gap. That's what we can provide a great software, but the U.S. they're really fussy, right? So when Microsoft first launched on their products, they actually went to Japan because they were really fussy customers and fussy customers build great products. So -- we look forward to -- and that's also why we spent a lot of time of enterprises because like they care and they're very detail-orientated and so that's how you build great products. So that's some locations. Then revenue per location. We're at EUR 570 million. We make roughly EUR 500 million from selling the practice management software, EUR 70 million from payments on AI today, right? There's a huge opportunity for payments, which is around EUR 450 million. And then we also have the potential to upsell AI. We're actually making it quite affordable for everyone. It's around EUR 120 million. And then conversations, which is something that they use third parties for which we're bringing in-house. And that's with the current pricing that we have, that the change that we can have in ARPU by upselling our customer base. So I talked a lot. I want to summarize. So what the takeaway? Number one, we are the #1 player in Europe, right? So around EUR 30 million ARR, right? We've got great unit economics. Like the reason I bought this business is this churn rate. It's a great -- we think to have very low churn rate and faithful customers. Second, we've got good net retention as our customers keep growing. We've got the opportunity to upsell them net new products. And it's relatively -- it's very affordable to recruit new customers. Like a good LTV to CAC is 3. We're at 27%, so. And then the third one is we don't need to go into new markets to be able to grow 19 times bigger, right? So -- and finish with the year opportunity, right? Now I hand it over to James, who will go through a bit more detail about the opportunity. Thanks for your time James, I'll hand over to you.

Okay. So just firstly, can everybody hear me? All good. So to put this into perspective, the last time we had a Capital Markets Day, which was in 2022, there was a new upstart product called ChatGPT that just launched. Like a lot has changed in the last 4 years in terms of what we expect out of software, what we expect out of everything that we do. This section is just meant to show you kind of where we are at the moment, how we've gotten there and where we're going in the future in terms of AI. I'm also very aware that AI is one of these empty buzzwords. I would say I've got some examples of things we're doing in the product right now. If you're watching this and you're physically in this location, we just come next door at lunch, we'll give you a demo of all the latest things and just show you how this is transforming is really exciting. So we've got a really interesting sweet spot that we find ourselves in. In that -- we've been around for a long time, and we have these things at the top, which are assets that take years to build. So being able to operate in many different locations, to have all of the workflows, the permissions, everything takes a lot of time and actually working with customers in real life. And we have also, over the time that we've been operating, we have structured data for all of our customers, thousands of clinics. We have a really rich amount of data behind the product. And probably the most importantly is that we, as a company, are highly trusted in our industry. we've got flagship customers, some of which are in the room, thank you for coming, who trust us to store their data, protect it and make sure that the right things are done with the -- we have that, and that's the foundation. But what we also have, which is really interesting is where we are today is that we have a huge amount of engineering velocity. We have a very strong team, which are getting to in one of the coming slides. But importantly is that in the last year or so that I've been here, we've really upskilled the team. We've completely changed how we deploy software. We've ripped out and rewritten large portions of the STACK so that fundamentally now, we're starting to manifest into the products, all of the AI things, which I'll show you some examples of from scribed to summarizations to Ask Provet. These things are now manifesting as a result of the last year of what we've been building. So this slide as well, if you see this kind of great thing going on. We kind of have the best of both worlds. If you're an incumbent PMS like ezyVet or one of the other capacities that we have, -- you are not this independent technology first SaaS company, which is what we are and what we intend to be. So yes, you may have depth and breadth of data, but you don't have the infrastructure to do what is needed for the future. And also, if you're a start-up PMS, yes, you may be coming to market with a new tech stack, but we have the trust and we have the operational capability across many geographies to do everything at scale. So we're really uniquely positioned in the middle. And also with these AI point solutions, which you see as well, there's lots of subscribers. There's lots of these nice little tools that you can use Sure. They're great. But fundamentally, if you think of like a hub-and-spoke model of a wheel, we are the hub. We are at the center of the clinic. We have the data. We have the knowledge and we have the trust. So we find ourselves in the middle of a really nice situation where we have both the data and the trust, but also the skills in order to really move us into an even stronger position with our software. We only really have to just add agents now on top of what we're doing, we don't have to rebuild everything from scratch. We've done that journey. What we've been doing over the last year is effectively a rebuild of the team. When I joined, it's safe to say that we were quite underutilizing AI, both in terms of how we build software, but also how it manifested in the product. That's been my focus since I got here. I joined just over a year ago and I used to work at Shopify somewhere that was very first to market with all of that kind of stuff. Like can we do this at Provet? Both myself and our new VP of Product, have been on this transformation journey with the team. It's an expectation now that every engineer who works for me, you are generating code. We're using AI tools in order to ship software, we've gone from shipping weekly to hourly. We've already shipped about 6 times today in the time that we've been awake. It's a big R&D transformation. On top of this as well, we are, at the moment, very, very fixated around the talent density of our department. I don't have any big plans to triple the size of the team. I want to keep the team the same size and then make sure that everyone who sits in one of those seats is both a high-performing individual and they've earned their place. I see over time us staying the same size, we're even getting slightly smaller because at the moment, with AI, as you know, like everyone's theoretical upper bound of their productivity is unknown. We're on this journey of seeing how productive and how efficient we can be, we're fixing the size of the team and then just seeing us grow that way. Just in terms of some sort of stats at the bottom there, 100% of engineers now are using AI tools to generate code, almost 100% of our code is generated now. We don't write code anymore. We've spent a huge amount of time working on our code base and our tooling to allow agents to work within it. Just measuring from when I joined the end of last year, for example, we had a 130% increase over double the amount of lines of code shipped to production and doing it basically hourly or less compared to what used to be monthly releases. Under the hood. Over the last year, there's been a huge amount of team rebuild going on. This also kind of transfers to how we think about building products as well. Amy, who I partner with I've worked with in the past at Brandwatch, which was one of, I think we'll probably remain to be in the future, the South Coast of England's biggest exit. We grew Brandwatch there and sold it to Cision. So high trust with my product partner and all engineers, all product managers, are super fixated on the two things at Charles that are really important, which is like that need to save time and also they need to run their businesses really well. And that's everything that we're focused on. In terms of the technical side of why that's important, we have completely rebuilt our data warehouse. So all of the data that Provet is now in a completely different system when I joined. We effectively have one clinical data model for all of our clinics all of our countries, all of our locations. The storage is far faster than it used to be. Not only is that good for scaling customer growth, all of the stories that we've rewritten everything into is built for AI. So when you use Ask Provet it's the data model that we have rebuilt that's enabled that to happen. In terms of our stack, it's not exciting anymore to say that you're in the cloud, but we're a modern cloud-native API-first company and all the technologies that we choose now are to enable AI and the products for our customers. The really important thing as well is that because we've been in the game for a long time, we have great compliance, great audit trails, all the kinds of things that are kind of after thoughts that the start-ups will have to think about. We have true scale. We are multi-region, we have a huge amount of data. The more customers we have, the more data we have the more insights and the more that we can do with that data as well. It's a virtuous cycle, which is really nice to be in the center of. You can talk about AI, but it's worth just to see in the product. So I think I took the screenshots on Friday. This is real. We've talked about AI scribe and we talked about summarization, which is where in the consultation, you can speak and it will capture the conversation, create the transcript, create the notes, create the AI actions. Again, if you're here in physical location, you can get a demo, but also by rebuilding on the back end, we've enabled something I asked Provet, which we recently shipped -- so the other part that we find that struggle with, especially if they're not part of an enterprise group. If they're a solo vet or if they're part of a small chain, even though they got into veterinary medicine, they unwittingly also have just become an entrepreneur even though maybe that wasn't the training or background. A lot of the challenges well is like how do you run a good business as a bit, which is hard because you have to think about a lot of different things. By using as Provet and putting agent on top of our new data stores, we can provide the whole kind of ChatGPT, Claude experience in the product to allow you to explore your financials to alert you to are there any gaps in your revenue? Is there any action you need to take? And as time is going on, we're adding more and more data to this so that effectively, it becomes your sidekick that helps you within the app in order to run your business as well. We're really excited about this. You can ask open questions, you can generate data for you in tables, natural language. It can give you suggested follow-ups. It can break things down in charts. Really, the whole experience that you expect by using Claude or ChatGPT or similar, we've brought that into the product, liberated on all of your data so that we can help vets run their businesses better. Again, about the foundation that we've been building over the last year rebuilding all of our back end to enable products like as Ask Provet, the nice thing is that within an extra 4 weeks, we shipped our MCP, we'll be doing the marketing launch that very shortly, where for users who love Claude or ChatGPT, they can just authorize with our MCP server - and then outside of Provet, they can ask whatever questions they like about their practice. As an example here, using Provet appointment data, make a graph of my busiest hours and sure which time I have the most customers. Using the MCP server, it can tell you that and then you can go and build your own tooling, you can charge it how you want, you can export the data. By getting those foundations in the right place over the last year, we're really seeing all that underwater iceberg start to kind of come above the surface. The fruits are sort of spilling out now, which is obviously. This kind of gives you the direction that we're going in terms of AI. It's been a rebuild of the team. It's been a rebuild of how we work, and it's now manifesting in the product. It's just the beginning really. There's a lot that we want to do. With that ends my section on AI. I'm really excited about the stuff that we're doing. I'll hand you over to Alex for the next steps, I think.

Thank you very, very much Charles and James for these incredible insights into the massive opportunity that we have in Provet and the product we're going to build in order to capture it. We're going to take a 10-minute pause now just to allow the fireside panel to get set up. And so for those watching the webcast, we will resume the session at 1:45 U.K. time. [Break]

We've got a fireside panel with two of our largest customers. One is Richard, who is CEO of Vets for Pets. And we've got Graham, who's Director of Innovation and Transformation at CVS, two of our largest partners. I want maybe to -- we'll have a discussion now together to talk about the industry to talk about the practice management software as a space and then their experience with Provet as well. So maybe I'll kick off with -- if you could briefly introduce yourselves and your business, maybe you can start for Richard.

Sure -- so yes, so I look after the Vet Group for Pets at Home. We are sort of the largest first opinion practice, which means essentially means an entry-level vet group in the U.K. We make up about 10% of the vet market in terms of the sector. And I've been with the business now for just 2 years. So my background has been broadly in operational and performance management, including military, operational support, retail and hospitality. So the vet industry is a relatively new one for me. I think one of the really interesting parts of it has been in recognizing how actually, there's a huge change, and we'll talk a little bit about the change in the sector in terms of how pets are seen, but also in terms of the brands that look after the clients that own pets and how important that is and how that's changing as well. We have 460 vets practices across the U.K. 330 of those are within the pet stores and then the remainder is stand-alone vets across our ecosystem. We have a very unique model. Our vets model is essentially very similar to a franchising model. It's called a partnership model. So we work in partnership with all of our vets practices. The difference between franchise. We provide all the service support that a franchisee would and also the management support as well to support all the management administration behind running a practice centrally. We -- but what we don't do is we don't dictate any operational or clinical guidance. We provide support. Our practices have complete operational and clinical freedom, and that's the difference between franchising and partnership. We have a really unique model, 460 practices, and we have about 600 different practice owners. When we talk about entrepreneurs, we have about 600 different variations of that all under one brand banner and all running local practices as part of a national chain, if you like. Thank you.

Yes, Graham Dodds. I'm a vet surgeon by background. And as Charles mentioned, I'm Director of Innovation and Transformation at CVS Vets. CVS Vets is a listed company. We've been around for I don't know, 20 years now, I think growth predominantly through acquisition, mostly in the U.K. We have about 500 locations across the U.K. and Australia, and the vast majority of that is first opinion, veterinary practices. There are some similarities. We don't have a tell approach with our practices. We very much support them to grow. We recently moved to the FTSE from AIM -- that was in March last year. I think although our results aren't produced until September, our interims kind of stated that we drove about EUR 700 million in revenue this year, which turned into EUR 140 million EBITDA. We employ 9,000 people. 3,500 of them are vets. It's a similar number of nurses, maybe slightly more. And yes, I've been there about 10 years myself. As I mentioned, I'm a vet by background before CVS, I had my own practices, grew them, sold then into CVS, which is how I headed up there. So very familiar with that journey. And in the first 5 years of that 10-year tenure ship with CVS, I was in operations. So I ran our referral division. We have a few multidisciplinary referral centers scattered across the U.K. And I also ran our Equine division as well, one of our areas of veterinary practice, we decided to diverse into a while back was Equine and I ran that. But throughout my whole career, I've been really interested in challenging the norm which is -- and I used to do that a lot at the exec and for my sins ended up in this role. For the last 5 years, I've been overseeing the transformation of many of our processes, but most of them center around technology. which is I assume why I find myself sat here today.

Maybe I'll start with you on the -- what you see as the biggest source of friction or inefficiency that led you to make a change in practice management software.

Yes. Well, I think that's directed to myself. -- yes. So I think when looking back at the systems we did have before, I think it was mentioned by one of your guys earlier that we had a standard kind of traditional practice management system software. -- that have written on Visual Basic 6, a very old piece of kit running on practice servers. I had to laugh when someone mentioned that earlier. We were that operator. We have made the decision to move to 1 PMS way back in the day for what's the reasons. But when we sat down 5, 6 years ago and thought about the future, and we thought about the things that we wanted to achieve. -- clearly, this wasn't the route for us. It wouldn't open them up. And we saw it at the time. I think there were three key things, I think, thinking back to that PMS that were a challenge at the time, let alone our future aspirations. And I suppose those three things. The first one was inefficiency in the console room. There were lots of clicks to get anything done. You've talked today about making life easier for the vet it certainly wasn't, and this was measurable. We saw it in things like billing audits. We will do billing audits to check that billing matches work done, and it didn't -- there was a lot of opportunity to kind of close that gap. Now not all of that gap was the PMS, but a significant portion of it was it was tricky to bill accurately was just one example. I suppose the second of the three, was as we grow as a business, we wanted to be able to standardize and centralize a lot of processes. I mean I've just mentioned we're there to support our practices not tell them what to do, but it's much easier when you've got SOPs and protocols to help do that. We couldn't do that. I mean, again, because of the on-prem infrastructure, I think we had about 450 instances of that PMS across 280 servers and to try and put anything out centrally was now on impossible. And I suppose the third thing was data, a big, big one. We really wanted to get more out of our data. And we've been working very hard to take data from hindsight to Foresight in CVS. -- hindsight, we were okay. Actually, we could pull reports from our last PMS, but they were pretty static and pretty slow. And again, due to that infrastructure, hard to get access to. So we needed to move. And so those are sort of three big things outside of cybersecurity and various other things that made us think we need to make a move.

What about you, Richard, do you chances to resonate with you? And what was the original decision for.

It was like a shot to the heart when I heard you talk about server-based businesses earlier on because we are still server-based but transitioning out of that. Yes, definitely, clearly, one of the things about advances in technology, digitization and everything around that is that at some point, you have to recognize the need to change and legacy systems start to hamper you rather than helping enable you. That's definitely -- and I think you go through a period of denial initially because it's a big change and an investment as well. I think when you look at the last 5 years within the sector as well, the sector has grown disproportionately in terms of its volumes. So you need to be able to increase your capacity, enable an increased capability in order to really make sure that you can continue to grow at pace with the sector as well. And so you need the right support, enablement to help you to deliver that successfully as well as that. So I think really, for us, a combination of all those things in recognizing how we need to change and really move from being a very analog-based business into digital and AI-enabled business actually in terms of positioning ourselves.

And so maybe starting in history, CVS was the first 1 to 2. So why did you choose Provet versus any other practice management software out there at that time.

As Richard said, it's a big decision. And there is a little Yes, exactly. Your churn is yes, important Yes, it was a big decision. Look, but we knew that we wanted to do that. And the examples I just gave are just -- they were today problems of the time that I didn't even touch on where we wanted to be in the future. So it was clear. That's what we needed to do. So first of all, I think we started off with some golden rules that we would we wanted to go out and explore all PMSs with and they included things like we wanted a cloud-native product, as I mentioned, cybersecurity already. We wanted something that we could integrate with other third-party industry-leading technologies as they developed to give us that flexibility and scope. And when we apply those golden rules to probably 30-plus PMSs we looked at, those roles alone whittled the list down to very few within which Provet sat. We then came up with an internal assessment tool that we designed ourselves really. And again, that looked at some of the problems that we have that were today problems. So things like diary management, billing, stock control, all that kind of bread and butter stuff. Beside that, some of the future things that we would like to be able to do. And we then scored that. We literally spent time with yourselves and other players. And you might remember those days, we got access to the sandbox and played with them all and we scored the literally scored them, and Provet consistently came up as the leader as a product. So the next thing, obviously, of course, all the due diligence and things needed done, but the next thing was for us to do a trial. That was to really sense check our conclusions on the product itself. But also we can tell you this now is to test whether or not you guys were going to be a good partner for us. And of course, it went well. I think we did one practice that we really focused on for a couple of months, and then we moved to half a dozen practices for a few more months. We were testing things like how well you guys could support us with the changes that we would expect to see, how adaptive and responsive you would be to things that we believed were gaps and stuff. Just general support, Richard mentioned is about the partnership for scale, the big scale business that cost is really important. That's how we did it, and that's how we ended up with Provet. So to summarize, we had a very open mind to begin with. We didn't have a view. And I think even back then, it wasn't so clear. It wasn't so clear how many of the players that were out there were really trending towards the top. Fast forward to now, I think as you've conveyed today, that's a slightly different picture. But yes.

I think we recognized really quickly when you look at -- so one of the things that Charles has brought out, which is relatively unique is that you're in it for the long term in your partnership. So not only are you making -- you're making essentially 1/4 of a century decision on the back of a change when you look in the rearview mirror, has not been made more often than half of that time period. So anything from 15 to 20 years. So it's a really significant moment in time. It's a significant change decision. There's a lot you have to consider within that. We did primarily the same. We took a representative group of our own practice owners because that was really important. Slightly different needs because you've got a need to have a certain level of local customization in order to have that clinical and operational freedom. But at the same time, within a framework of a really sound and solid future-proof system with a partner that you know is going to be there in the next quarter for Century. That's really important. I mean, because it's a substantial investment as well, of course. Alongside that, we reviewed and tested 50 different operation systems and then slowly work those down. I think partly, I would reflect, I think the sales team was one of the best definitely with what it promised and actually some of the stuff you are delivering now. So -- that was pretty true. But not only that, yes, I'd echo what Graham said about partnership that you're trying -- there are two things here. An enterprise system with international influence is really valuable because you get to gain all that benefit as the system develops and the inputs internationally are also helping to develop the capability as well as us also contributing to that, too. A really meaningful international reach enterprise system has been really important to us. Then alongside that, I think just a great partnership, where there is always going to be give and take and you don't sweat the small stuff together because you really tackle the big stuff, you have patients with each other, you understand each other as much as you can, and you challenge each other as well in the right ways. And I think that I'd always measure the value of the partnership and how the team feels about actually having -- interacting and working with. And if the team enjoy it, then that's a really good sign that you've got a good partner.

And maybe for both of you, like you had your structured assessment, you did your pilots. What came out after this that you didn't think was important, which you think is important today?

Yes. There's one for me which I kind of noticed and we'll probably come to the rollout that we did later in the conversation. We -- I don't think we assigned enough importance to how intuitive the system is, actually. We -- I don't know if we kind of overlooked it or not. It seems obvious now with hindsight. I think we were just so used to platforms not being intuitive. And we -- in CVS, we've got, we call it, Knowledge Hub, sure you'll have similar to training platform. where we have a mixture of sort of bespoke training that we've done for whatever we need to help support practices with. The practice management systems section and that was huge at the time for our previous system. And we did -- you may recall, we did -- we came up with some videos and training sessions for Provet. We did it ahead of our ultimate rule out we did -- but we always are monitoring everything. And we were looking at compliance of that training versus engagement in some of the workflows. And again, there was a huge mismatch. There wasn't much training done. The team was managed to do it. Our conclusion was they either taught themselves or they learn from each other. Both of them are great wins in the spirit of working together and getting everybody on the same page. As it transpires, as I say, I'm sure we'll come to it. We did quite a quick rollout of the system. I don't think that actually would have been achievable if the system wasn't so intuitive. So I suppose that's one thing I would acknowledge.

Yes. I think the two things for us. The first is the realization that you're -- so you're having to engage with a group of practices who aren't asking for this change. They know it's necessary, but they haven't invited it. So you are forcing it on them because it's necessary. You come to end of life with one system. You've got to move with the times. And so the recognition that there have been over a decade of customization on something they felt really comfortable about to say we're not going to change that and help you to understand the reason for the change, but not only that, actually, really them help them to understand the intuitive nature of what a new system might feel like. I think alongside that, and I think this has been probably the most interesting development is, I think prior operating systems were highly administrative in terms of how they support the way the practices work. The reality is now that the development and capability of the platform that you run in a vet practice actually starts to inform you about how you change the design of how the practice works and how the practice delivers its experience to its clients as well. There are probably three key areas there, if you -- and I know some of these have been touched on. So the first is being able to consult with the client and their pets without needing to touch a keyboard. That's a massive change in shift in terms of care interaction and also the capability of being able to then harness data to be able to prompt and suggest and really enrich that consultation, which can all be fed -- can actually be fed through that whole experience. Then removing time in terms of actually then having to make notes afterwards and how you administrate that. The connection between the consultation room and then the administration within the practice and how that can be enabled and completely changed and differentiate in terms of providing different levels of information and also providing information in a really CMA compliant way as well. So one of the key things from the Competition and Markets Authority has been about transparency and about itemization and about clarity. Actually having a platform that enables you to automate all of that and provide that in the right formats actually means that those elements aren't things that practices have to think about because they can happen automatically. Then the final part, which is one that has surprised us is, there are three things that come out really clearly from clients is customers. The first thing is that -- the #1 thing that they value actually in terms of digital integration is online booking and how useful that is. Even if you've got a terrible online booking system, they value it. But actually, AI means that you remove clicking, any clicks to an online booking system and it becomes a conversation about then customizing your -- the booking to the best slot your best preference. When you then step back from that and you think about the client experience in a reception because the one thing about vets is you're always going to get physical visits, yes. That's guaranteed. That will always happen. But actually, the experience within a reception in terms of Am I acknowledged by the reception team is my consultation on time, how well informed am I currently sits around a checkpoint, which is the reception and actually being able to remove the reception desk and make it more about a client hosting experience because the digital functionality of the platform can support more self-help from the clients and more guidance, suddenly revolutionizes the way that the client feels about. First of all, experience is care and then how they feel when they walk away from the practice and think about your brand or about the experience. And those things have really -- I mean, I think they've been certainly more recent beneficial developments, then they will just become more and more relevant to what we do.

Maybe shifting on to rollout. So in 2024, CVS rolled out almost 350 clinics in 6 weeks. -- still remember the call for the 8-week deadline back in the day. So what made it possible to implement that speed? And what did you learn from that experience? If you have we do it, how would you do it?

Made it possible -- it was lots of sleepless nights. This -- the product helped a bit, but the people clearly were really important, having the infrastructure around us to do that. it wasn't just rolling out the PMS for us. We also had to take care of hardware and various other things. So there was lots of stuff going on in the background. Within our business, we had project management support, change management support. We also built a team of specifically trained PMS support people, which you helped us with. You provided your own, and you trained up some of our of our own people. Yes, a lot of late nights. -- having trust in Nordhealth as a partner was key. We really had to make sure that data migrations and things were done every night on time. And I remember, practices would close at 8 p.m. The data migration would be done, tidied up, checked. By midday, the next day, they would be all on their new system with the new computers. But that was it, really making sure we had the right people. There was a lot of work done around explaining the why. As Richard has kind of covered already, it's a big change, and it's -- they do people in practice feel like it's been done to them. But if you explain the why, which I think we did effectively, it worked. Moving at pace was, I think, a good thing because quite quickly, it creates a critical mass and everybody was in it together. And I think once we've made our minds up after that trial, that's what we wanted to do. it was the right approach. Whether I would do it in 6 weeks again, I'm not sure we wouldn't go that fast. But certainly, I think once you made your mind that going for, it's the right thing.

And what about Richard, so the 50th clinic is now on proved as of July right? And how is the experience with the implementation of so far?

Well, we're about through 75 now. We'll be 100 in September. For us, I think it's always the fear of the unknown, I think, is what prevents you from feeling comfortable taking certain levels of risk. And I really admire actually what Graham and the team have done. And I think there is always a great traction to actually leaning in that direction and going, you know what, actually -- there may be some short-term pain, but getting everyone through it and altogether on it actually is not necessarily a bad thing. I think -- so we have taken a I think a more considered approach simply because we are very conscious of each individual practice being business owned essentially. That doesn't necessarily mean that we won't consider and shouldn't change the approach we have, but we're certainly at a pace now where we're doing 12 a week on a regular basis and continue to roll through. We're seeing a much better level of consistency and confidence, which I think is really important, and that starts to then build advocacy and then build understanding and build excitement actually, to a certain extent. Yes, I think we're in a good place. I think you have definitely challenged our thinking more than sometimes we'd like to but I think that's been -- that's what a healthy partnership all means. That's really helped us to also, I think, hold our nerve when you make decisions to step into the unknown and rely on that experience that sits around you to give you confidence that you're not stepping into the unknown alone, you've got experience and support around you with that. We did have rollback and contingency in place. But I think of those 30, I think there was only 1 or 2 that didn't quite go to plan, and we got there in the end. But yes, it was interesting times.

And so now shifting to AI. Like where do you see the AI creating the most immediate, measurable value for your practices?

Yes. So I think probably touched on the three key touch points that we've identified. But actually, the reality is that it's only as good as the data that you provided with -- so for us, actually -- and it's one of the big benefits of cloud storage, again, is actually data sharing is much easier and smarter. But actually, also, we've got to remember that we're part of a much larger pet group. And within that, we've got an incredibly rich seam of data in terms of our clients, customers and on everything to do with pet care actually -- so really being able to harness that not just because it's beneficial commercially for the group in terms of actually being the U.K.'s only complete pet care provider. But actually within that is a vet group, being able to use what is essentially independently built data at a server level by practice to be collectively harnessed for the -- as a total means that every practice benefits from that and benefits from the information that, that provides to help it to make decisions to be more intuitive, to be more efficient, more effective and to help guide it based on really good quality information. Without the data, you can't harness it. I think for us, that's been really important in recognizing how significant that is because that sets the foundation. And then you've got to enable it with the with the way that you can provide data and share it in order to then be able to really harness it.

What about that CVS, where do you see the AI opportunity to transform your workflow.

Yes. Well, I mean I've got to agree with everything Richard has just said. I think going into the consult room and looking at some of the stuff that you can showcase today next door for the people in the room, really clever stuff. And that I do think that the opportunity for efficiency lives mostly in there. Any interaction around the customer or client pet parent to use your terminology, to give that vet focus and time to interact with them properly. I see a huge benefit there because it does. As I say, as a vet myself, it does distract you somewhat having to constantly turn away and make sure that your notes are typed up when you're already running 5 minutes late for your next consultation, et cetera, et cetera, you get the picture. But then also the client experience that we would love to deliver that for the same reasons we struggled to do. I mean I think it's a -- I don't know if it's a real stat, but it said that a veterinary customer will leave the consult retaining only 5% of the conversation that it's had. We'd love to be able to send them home with access to comprehensive notes that are relevant to them. That's easy to do when you've got plenty of time, but you don't whereas AI can do that. And as you know, we've been trialing that element of your product for a while now, and that's really transforming that customer experience as well. I think the sticky touch points are in the consult room and giving the client experience that sort of extends beyond that, I really see a huge value there in the day-to-day workflow. But then within the data as well, I mean, again, Richard touched on data already. I mentioned earlier, we want to get to the point of having real foresight and understanding what our clients are doing across the business. CVS as well, we do have our veterinary practices, but we also have our online pharmacy. It would be great to have a single customer view across our entire estate and start to understand what behaviors if anything, do exist out there even if it's down to the disease, what is the incidence of arthritis at certain age groups and certain breeds what might that do for our business? Should we be reaching out and utilizing this platform and CRM to contact customers sooner with breeds and species that fall into that bracket, and get those patients into care faster and sooner, which is great for the pet, great for the owner and is hopefully commercially successful as well. So two key areas, there's the sticky touch points that make the workflows clunky. Then I think there's a huge bit on data where we could start to strategically inform how best to grow the business and provide better care to those patients that are going to become ill as they get older.

And how do you make the decision on whether to build your own, right? But you have access to development resources or by a third party for that or have it in the PMS.

I don't know it's early thinking this one. We've done a few -- we've trialed third-party apps at the minute across our estate. They've gone well. They're good. some of them. I think what you're doing now is I can see the acceleration, and we're utilizing your product really well. I feel -- my gut feel at this point is that the more you can have within the PMS probably the better. I'm happy to be challenged on that view. But I think although at the minute, the use cases are early. I think as things expand, applying the AI beyond just scribed technology and extending it into looking at the history of that patient in conjunction with other anonymized data from other patients with same backgrounds and things like that, I think it would be much faster, whereas third-party applications don't have that level of integration. It can't work that fast. And then there's just a simple administration of that technology in the consulting group. -- having more windows open is more clicks, which is clunkier, et cetera, et cetera. I got a feel from the way I see it evolving. It's nascent, but I think if that can be achieved, which it looks like you're on the track to do, that would be my preference at this point. I'll keep an open mind.

I'd agree, simple is better. I mean it's the constant development in technology and how you can use it. If it doesn't make things simpler and easier then it becomes a real challenge. And these are highly complex systems in terms of what they're able to do, but you don't want to have to worry about that. The complexity of the system should take care of itself. Actually, the simplicity of how that means to reducing actually the number of providers you work with reducing the number of systems you have to try and access in order to enable you, I think, is always going to win -- and I think simplicity is such a critical part of being able to deliver success in another complex world that that's it.

Yes. We both want our vets and nurses to vet and nurse the more we ask them to start playing with multiple systems to achieve an outcome, the harder that will be and the compliance will drop. So I think that fits with that conclusion there.

Maybe -- We're very deep on the veterinary industry, but just for the investors here. Could you -- if we zoom out, like we've had the pandemic make a boom in pet ownership. So what do you see the outlook for the veterinary industry in the U.K. more broadly in terms of like a couple of years?

I think -- so yes, what's been fascinating. So not only have you had a massive growth in volume of pets, 38 million pets within the U.K. You've also had significant capacity and growth within the sector as a result of that. Pet ownership has changed significantly, too. We -- the U.S. commonly refers to pet parenting pet parents are a very clear definition of how pet ownership has changed. The dog in the home now has a place on the sofa in front of the TV. There used to be in a kennel the indoor plant in the sitting room being replaced with the cat tree. And I was in -- was one of our pet care centers the other day. It was very funny. And there was a couple there in a dog aisle and one of them brought out this dog safari outfit and said, what do you think? And the partner said, that's perfect. We need to get that yes. So those -- I mean that just again helps to reinforce how the pet sector has changed significantly. If you zoom that into what that means for vets, I think that the sector is going to continue to be -- remain strong. You've got a very high volume of Covid pets who are reaching maturity and going through maturity in the next 5 years. We refer to that as a smile curve in terms of the maturity and the need for vet support. I think that we -- you can also see that technology continues to move on. The industry is really shifting from one of treatment and responsive care to preventative care and all based around well-being provision through life stage. That's a big shift and that will continue to move forward, particularly pet health, preventative health and how insurance and subscriptions wrap around that to provide and make pet care much easier.

Yes. Just literally the same as I was thinking of better mentioned preventive pet care than you did it already. Yes, I walked around Kensington Gardens earlier and there was a few dogs in prams, it's interesting to see. I completely agree. But yes, I think the future is really bright. I think the post-covid boom, as you say, veterinary spend is somewhat reluctant when it comes from a client's perspective when it comes to reactive care. There's no doubt that the puppy and kit and boom that we saw during COVID, that will come. I think, as I mentioned earlier, utilizing technology to make sure that we get that opportunity to speak to those clients as soon as possible is going to be key. And as I say, using data and using CRM, I think we'll be able to do that better than we've ever done before. I do think that the way the generations are changing and our pet owner ages and the Gen Zers that are coming through that on these pets, they want to be able to do things in their own terms and their own times -- they belong to a subscription model world with Netflix, Prime and all these other things. Our health plans that we've had in the sector, which in the past have been product-led to cover the cost of worming, fleaing and vaccinations and things like that will and should change, I think, into service-led propositions, which will enable, I think, clients to bond with their vet throughout their pet's life. -- so that when the pet does start to become ill, they can be seen sooner and that reluctant purchase that I mentioned will be somewhat less reluctant at that time. So I think with tech enablement and those changes in dynamics, I think the future is very bright for the sector and I'm very proud to be working in.

Thank you -- thank you both for your time and thanks for your partnership. Thank you. Graham and Richard, thank you so much for coming down today and doing the panel with Charles. I think you've given our investors and an audience more insight into our -- into the industry that we serve, the client base than we ever could have. So a massive thank you there. We're going to take another comfort pause now. say, 20 minutes. So we will reconvene at 20 to 3 U.K. time, where we'll kick off with Karan on the therapy business. So see you in 20 minutes. [Break]

Welcome back, everyone. Next up, we have Karan Wallia, our CEO of the therapy business, who will be taking you through the strategy for that business.

Okay. Good afternoon, everyone. I'm Karan Wallia, CEO for the Therapy business here at Nordhealth. I spent a fair amount of my time with customers. So that's therapists and clinicians. And -- because I think that's still the fastest way to learn about our business, the market regulation trends, our customers' business. And it's often you pretty much hear the same thing, which is these therapists, they've been trained for years to serve patients and treat patients. And still, they have to spend their evenings, nights, weekends, doing hours of documentation, filling forms and writing notes. And for every hour of administration that we can help them save is an hour they spend towards treating more patients or providing better care outcomes. And that's really the job. That's what the therapy business is about. That's what our mission is and what our product solves for. Over the last almost 2 years that I've been here, are focused on, yes, maybe a couple of things. The first is enabling the migration of our users in Norway. The second is consolidating our lead in Finland and lastly, building an AI native operating system for therapists in the Nordics. So our core segments are psychologists and physiotherapists, so musculoskeletal and mental health. And -- the first thing in the morning, they open our software. They spend their entire day in it. So it takes care of everything from patient intake, their calendar availability, medical records, journals, invoicing insurance reimbursements and so on. So really the administrative side of practice management, along with medical records, electronic health records in one system. And -- so our system is sort of the center of gravity for the clinic's data and their operations. And we have 13,000 clinics, 27,000 therapists that trust us with the most important part of the clinic. The most difficult things to move for our practical purposes, which is the clinic's memory, so their data again, and their daily operations. So unlike a point solution like Telehealth or Patient Engagement Tools or even an AI scribe. -- we're not optional. We're not an add-on tool. We're really what the clinic is built on. So they build the entire practice around it. They have 15 years of patient history and therapist history in that system. So patients and therapists will come and go, but really, their practice is built on the system. And I think that's what makes us quite defensible. Also, the third point here, compliance. That's, generally speaking, quite an underestimated factor in our business. Every time regulation increases, rules tighten or data privacy rules tighten in different markets. Audits which are becoming more routine certifications, which are becoming more expected to win larger accounts. This continues to increase and every time this increases, it makes it a little bit harder for a new entrant to enter this market because the barriers continue to go up. But for us, this is business as usual. So organically, that gap against potential new entrants tends to widen for us. When you take that along with the trust that our customers give us the churn rates that you can see the scale where we have the opportunity to serve more than half the therapists in our market along with the R&D muscle we have, especially around our AI capabilities. That positions us very well to win the race to be the system of action as Charles calls it or the agentic PMS. And when we think about the Agentic PMS, agents are only as useful as the data they have access to, right, or the context they can work in. So really patient history or for example, the regulatory protocols that you've got to follow in different markets, although access to workflows you can act on. That's really what differentiates us when we think about this. Also, I think being early on this path is definitely helping us widen our gap from existing competitors in this market. Also, when we think about the opportunity ahead of us, we serve a large part of the market in terms of the number of therapists. But in terms of the total TAM, there's still some distance ahead of us. We're only 28% of the total TAM which we look at as the number of therapies and the available spend per therapist. We do a good job with taking our PMS to market, but there's clearly an opportunity for us to capture more of their wallet share. Doing that at margins that we've proven at maturity of scale, 45% cash EBITDA margins makes it very attractive. That really makes the opportunity very valuable for us. I thought it would be nice to share a little bit about our product and where it's headed. So maybe the easiest way to consume this is if you think about an afternoon or a physiotherapist before a patient comes in, there's already a patient summary with the important markers, follow-ups from the last session and so on to ask. When you're sitting in a session, our scribe is open, listening, transcribing so by the time -- before the patient leaves, the notes already written. Then our Co-Pilot is putting together documentation, whether that's a discharge note, treatment summary, writing a referral to another doctor or an epicrisis. For example, in Finland, there's a KLR reimbursement form, which takes 20 minutes to fill in because you've got to look at the entire patient's history before you fill it in to get the reimbursement. That's now two clicks away, one to generate the summary and the other to submit it. We're really having impact there with our copilot. You can imagine in the future, we'd also have like a reception agent that based on the consultation is already scheduling your next appointment, right? So one thing to notice is all of this is sitting within our product, right? It's not a separate tool. Like today, you have a lot of AI scrap that do a good job of taking notes and transcribing and summarizing those notes. But this is all within our product, right? So it has access to that context that I was talking about. Yes, this is why we think that we're in a good -- we're well positioned to actually move forward on this -- to work towards the system of action. Also a little bit about our market. Today, we're around EUR 17.7 million in ARR, and we see there's a lot of opportunity, especially to capture more wallet spend from our existing therapists in our existing markets. And we've always shown you the therapy business as a complete portfolio. But today, we can actually take a click down and look at Norway, Finland and Denmark independently where we have strong market share in Norway, for example, but the opportunity there is really executing this migration well because that allows us the opportunity to then have much stronger margins, very attractive margins. But also that then gives us the base to be able to upsell AI products or whatever other products on, which today on a legacy base is not possible for us. Finland, I see this more as a proof point for what a mature market looks like in our portfolio. So I'll come back to this in a couple of minutes. Then Denmark, this is where we have a lot more white space to grow. We've done a good job on the mental health side, there's work for us to do on the physical health side in Denmark. I think the key takeaway here is we still have clear headroom to grow in our existing markets. Also, I think the economics for our business are fairly durable. We have high market shares. We're still able to generate organic growth. If we look at our net retention and 4.5% gross churn, that's a signal that our base -- they stick around, they grow a little bit more each year. They spend a little bit more with us each year. Every euro we spend to acquire a customer, we roughly get EUR 15 back at maturity in Finland, for example, we have a 45% cash profitability margin. Markets that we choose to compete in, we, over a period of time, are the market leader. On the revenue side, we've been able to grow steadily, even though we're in markets with high market share. But underneath this, all our focus, effort and investment has gone towards executing a large-scale migration in Norway. The goal for doing that is to be able to actually optimize for long-term growth to set the foundations to then be able to actually expand our business with those users because it's a large installed base with which today, it's difficult for us to expand because they're sitting on a legacy platform. Instead of squeezing short-term growth from a legacy platform, we've chosen to take the long-term investment option here. Also, you can see our customer concentration. It's pretty healthy, which means we have a pretty diversified base, thousands of small durable relationships, which in general is hard to replicate. The same story shows up in our margin profile, where we've taken -- our margins have taken a temporary dip because of investment in two key areas. One is the migration and the second is building out our AI capabilities. What we do know is that this is a temporary dip. We do think the underlying business is healthy. As this deliberate investment starts to normalize, we can start to see the margins that we plan for actually show up as they do in Finland already, for example, which is much further along on the maturity curve. Talking about that, so Finland, we have a 45% EBITDA CapEx margin, but an easy way to think of this is, let's say, if any of you own this business outright, right? EUR 0.45 of every euro would turn into cash, show up in your pocket every year. That's basically this business. When I think about this business, it's a business that returns roughly half its revenue back in cash, funds its own after funding its own growth and doesn't ask for more capital to keep doing so, right? And these cash flows are pretty defensible, repeatable. We have 95% of our customers come back every year. This money tends to show up. And Finland is not special. It's just further along on the maturity curve that's gone through the investment cycle. Our goal is to replicate similar economics in Norway. As a business, we have two main missions. The first is migrating our Norwegian customer base from a legacy platform to our modern platform and second is AI. So on the first mission, our margins in Norway have clearly been held down by a fairly complex and migration that's taken longer than we would like. We do see that post migration when investment does normalize, we can be at some fairly attractive margins and aim for something close to where Finland is today. The levers to getting there are really our R&D investment tapering off us being able to retire some of the legacy license costs and operational synergies coming from being on a single platform. Earlier today, I got the question on, hey, is the migration project on track. The answer is increasingly, yes, right? We have 1,000 migrated users, we have 700 users that are already booked for future migrations. That number is steadily increasing. We start to see a more predictable pace over here. Maybe it makes sense to also share with you how we internally think about this project, right. Going back to first principles, these are not customers who are trying to win. These are customers that we already have. We know what the future looks like as Finland shows us what a sticky, high-margin cash generative business can look like? So it's quite important for us to be able to handle this migration in a very customer-friendly manner. I guess from the outside, cautious can obviously look like slow, but we do optimize more for lifetime value than for migration speed. Also, by -- I think by the end of 2027 we should be in a place where the vast majority of the migrations are behind us. And then we start to slowly see some of the some of the savings actually coming in. The second mission is around AI. I think I remember reading somewhere the best type of sale you can make is to a customer you already have. And that's really what this is. We have a customer who buys our PMS, we upsell them our AI product and they're worth roughly doubled to us. We have 1,400 to 1,500 paying customers today. They refer other customers to us. And so the nice thing about this is the opportunity is really ahead of us. We're early, it's nascent, and we have an opportunity ahead of us over here. Also, just from an industry perspective, the supply of therapists is not increasing. It's physically constrained. The number of graduates is pretty static. But the demand is far outstripping the supply for mental health services or even for our physical health services with an aging population and so on, especially in our markets. The only way to do this is to expand their capacity. That's really where our AI tool sort of comes in and our AI products come in. And -- as our customers grow, we have an opportunity to actually grow with them. It's interesting that the more pressure that there is on a clinic, the more valuable our product becomes, and that gives us an opportunity to grow as well. Finally, so everything we've spoken about so far has been very focused on Norway, Denmark, Finland, basically our Nordic markets where we still have some headroom to grow and opportunity to expand our ARPU. But when you take this in perspective of a global opportunity, we're essentially a rounding error, right? The important thing that gives us some level of conviction, some level of belief that we could mean something in a broader market in the future once we finished executing on our current priorities is -- we've been able to build a winning team and a winning product in the hard-ish markets in the world. Like the Nordic markets we operate in have are the most demanding in terms of regulatory infrastructure and compliance most demanding in terms of localization. They have the most change averse customers. So being able to actually win in a very, very hard market gives us a fair bit of confidence that we could replicate and leverage our existing muscle in the next set of markets that we might look at after we're done with our existing priorities. Then finally, just to wrap up, we're a clear leader in our category and our markets. That gap between us and other players as well as potential entrance is widening as we continue to invest rapidly on AI. Our market leadership position gives us the opportunity to have healthy recurring revenues, an opportunity to expand on an already large installed base that we have. We have a healthy churn. And at maturity, some highly exceptional cash generative opportunities margins as well over here. Finally, a lot of the opportunity we have at least in front of us right now with our current priority sits in our markets. And beyond that, we'll then do a strategic review once we are well underway on our existing priorities and then consider what's next for us in terms of our growth. So that's it from the therapy business. I'm happy to take some questions when we do our Q&A. Thank you all very much.

Thank you very much, Karan. We're going to turn over now to financials and guidance. I'm going to start by giving you all a quick update on Q2's performance for both business units. Starting with the veterinary business unit. In H1, we have signed three more 100-plus clinic enterprise groups. These three are currently all either in pilot, co-development or early implementation. On top of that, we've signed EUR 564,000 of new ARR in Q2 alone. In July, we also announced that we'd completed early implementation of Vets for Pets, which means that we're now on to the core rollout and rolling out in earnest. We're also very excited that our second clinic in Germany is now on Provet, and we hope to have many more in Germany onto profit soon. On our mission to build the leading AI PMS. Ask Provet went live in Q2. Our clinical AI agent went live in Q2, and the Provet MCP went live in Q2. And -- so you can see, as James explained to all of us, the pace of rollout is and pace of deployment is just increasing massively at Nordhealth. Finally, on our quest to reduce time to value. We built some migration tooling that we launched in Q1, and we've seen this cut onboarding time by more than 1/4 across the business, which is fantastic progress. one of the legacy platforms that we acquired in Norway, Sanimalis, we sunset in Q2, 7 years after we started the migration. So we're extremely happy to have the final clients migrated over from Sanimalis onto Provet. VetVision, which is a Danish company that we acquired, we are aiming to sunset by the end of the year. Looking at the veterinary ARR growth. We've been reaccelerating our year-on-year growth and ARR has grown by 13.6% between Q2 last year and Q2 this year. This is driven by a net retention rate of 109.9%, which is itself largely driven by enterprise clients, rolling out new clinics. As Charles mentioned, we have an incredibly low churn of 2.2%. The increase that you'll have seen since last quarter in our signed not implemented is now that we've completed the early implementation of Vets for Pets we've included the full rollout number in our signed but not implemented number. On the adjusted EBITDA minus CapEx, you'll see that in H1 2026, we are lower than we were in H1 2025. This is, as you're all aware, due to all the increasing investments that we made during 2025, particularly on DACH localization and AI transformation. While we are on a path of increasing profitability in the vet business unit for H1 versus H1 last year, we are still a little bit below. Moving on to Therapy. Updates for Q2. Similar to the vet business units, we've shipped a lot on our path to becoming the leading AI PMS. We shipped the AI patient overview. We shipped AI-generated Epikrisis. We shipped custom AI document templates, and we shipped custom feedback templates. As Karan mentioned on the Aspit migration, we have 1,000 users migrated as of the end of Q2, with many, many more booked. We have 700 on top of that booked to be migrated. We launched Medipay integration for collections, and we've done appointment overviews and the ability to create a managed group sessions. We've launched that as well. Then finally, on Therapy growth, there's been EUR 1.2 million of ARR signed in H1 2026 across new business and upsell which is up 24% on H1 2025. So we're seeing an acceleration of growth in the veteran -- in the Therapy Business unit. And you can see that in the ARR numbers, which are at 10.8% growth year-on-year, driven again by a healthy net retention of 104.7% this has been helped a lot by the AI scribe, which has been growing very rapidly in the Therapy business and again, healthy gross churn below 5% at 4%, 4.5% in the Therapy business. Therapy adjusted EBITDA minus CapEx has already been showing its curve back in 2025 we expanded the team and invested in the migrations. But now that we have a well formed team, we've been allowing that growth to reaccelerate profitability. And so year-on-year, EBITDA minus CapEx has increased in H1 from EUR 0.7 million to EUR 1.2 million in the Therapy business. I'm going to take you now through the group-wide financial updates. On a group basis, in Q2 year-on-year, we've grown 12.5% in implemented ARR as you know, driven on the higher end by the veterinary business unit, but with therapy growing at an increased pace as well 108% net retention rate again, was the largest amount driven by Provet expansion amongst existing large clients and the blended churn rate incredibly low 3.1% across the group. As I mentioned in the vet section, the increase in signed not implemented coming from that inclusion of the full Vet for Pets rollout. Now turning to our accounting quarterly reported recurring revenues. Q2 year-on-year reported recurring revenues have grown by 10.2% and up to EUR 14.2 million in Q2. Our underlying recurring revenue growth, which is the 1 we really care about, has grown by 14.6% year-on-year from EUR 11.3 million, up to EUR 13 million and that share of recurring revenue as a share of the total revenue, very healthy at 91.2%, increasing year-on-year. For H1 as a whole, we have had a 9.1% increase in total revenue, up to EUR 27.6 million. But again, that underlying recurring revenue has grown by 13%, up to EUR 25.2 million up from EUR 22.3 million in H1 last year. And again, a very healthy share of recurring revenue at 91.3% of total revenues. Looking at quarterly adjusted EBITDA minus CapEx for a while now, we've been saying that we'd be holding our costs relatively flatter and allowing that growth to filter to improving EBITDA minus CapEx, and you can see that in our Q2 results versus last year, we were at minus EUR 1 million for the quarter. Adjusted EBITDA minus CapEx. We're now at negative EUR 0.7 million. We're starting to see that -- those improvements in profitability filtering through. It was minus EUR 1.5 million last quarter for another reference point. In H1, EBITDA minus CapEx, you can see that for the H1 as a whole, when you add the 2 quarters together, we're still at minus EUR 2.2 million, which is a reduction versus H1 of last year. But as I mentioned, that trajectory is positive. And so we're expecting improved profitability as the year continues. How that translates to cash. So we have in our adjusted free cash flow, we have reduced by EUR 0.9 million. Our adjusted net result is plus EUR 0.2 million. We have had some timing differences in the cash flow. Actually, July was a pretty good month for collections. We're expecting some of these numbers to filter through, but this is mostly timing difference versus the EBITDA minus CapEx result that I presented for Q2. On the H1 adjusted cash flow. The adjusted net result for the full half year, minus EUR 0.4 million adverse, which is driving some of that EUR 2.3 million adverse. But another point to note is that in Q1 2025, we had collected a very large one-off backlog which is somewhat skewing this year-on-year difference by about EUR 1.1 million. The remaining working capital variances are coming from Q2 that I explained on the previous slide. Then finally, on our balance sheet. We retain a very healthy balance sheet. We have EUR 12.5 million of cash as at the end of Q2. We still have no interest-bearing debt. We have a very healthy equity balance of EUR 55 million. There's been no material equity transactions in the quarter and any changes in goodwill has been to amortization or FX. Turning on to guidance. For 2026, I'm very pleased to tell you that we are still on track for the guidance that we gave. We're not restating or reiterating guidance. We've said we're going to do a full year recurring revenue of between EUR 50 million and EUR 53 million. You can see from H1's actuals of EUR 25.2 million that if you annualize just that we're already on track, but we're continuing to grow through the year. We're very comfortable on our recurring revenue range and guidance. On adjusted EBITDA minus CapEx, as we saw, we're at minus EUR 2.2 million on our H1 2026 actual. But with that improving profitability through the year, we again feel comfortable that we will remain within this guidance range. Finally, on the 3-year guidance to 2029. You can see here that in the last few years, on a group basis, we've had year-on-year recurring revenue growth between 13.6% in 2025, up to 21.5% in 2024. That's our historic range. It's been a bit lower on the therapy side between 4.4% and 7.5%, a bit higher on the veterinary side between 21.5% and 35.4%. For 2026, we've been guiding between 9.5% growth and a 16.1% growth. I'm very pleased to say that for 2027 to 2029, we believe we can maintain a CAGR across those 3 years. Actually a slightly top end, even slightly elevated versus 2025 and 2026 range of between 13% and 17%. That's a 15% CAGR and from 2027 to 2029, plus or minus 2%. The reason we have confidence in this level of growth is across both business units. We have a very positive outlook. On the veterinary side, as I hope Charles has proven today, we have a strong right to win in the growth markets that we already serve. We have the most enterprise-ready solution anywhere globally. And we have a large upsell headroom on our revenue per location. From the veterinary side, we're guiding 20% and CAGR plus or minus 3% over that 2027 to 2029 period. On the therapy side, we've been demonstrating strong AI upsell across our existing base and across the future migrated base. We're going to continue to gain market share in the Nordics. And then post migration, we believe we can accelerate growth even further by investing in new markets and new product development. We've already seen the growth rates in therapy increase from these single-digit percentages that you see in the last few years to ARR growth that you've seen year-on-year today. For therapy, we are guiding a 10% CAGR from 2027 to 2029 with plus or minus 2%. Next, it is very important for us to be masters of our own destiny. We will, from 2027 be both adjusted EBITDA minus CapEx profitable and cash flow positive. Given the current macro uncertainty in financial markets, we feel that we have a lot of strength by really generating our own cash and using our own cash to fund growth and not having to go out for external financing at any point. If we do go out for external financing, it would really be for an M&A or some inorganic growth. But for our own operations and the numbers you see here, we are going to self-fund that through generated cash and profits. The reason we have a lot of confidence that we can become profitable in 2027. Firstly, we're going to maintain our healthy growth rates. Year-on-year recurring revenue increases in 2026, is between 10% and 16%, and 2027 to 2029. It will be around that 15% mark. We're going to let that flow through to the bottom line using our healthy gross margins at which in the last 12 months have been 78%. We still maintain very healthy SaaS gross margins that allow that revenue to turn into profit. On the sales and marketing side and CAC improvements, we're very disciplined in how we manage our sales and marketing spent in order to ensure that the money we do spend pays back within a reasonable time frame and is very profitable. You've seen from our unit economics that we have a lot of headroom on our CAC and that it's a very efficient use of our money. We will leverage that in order to retain profitability. As I mentioned earlier, our onboarding efficiency has increased as well. And so we've reduced time by 25%. We're going to continue to improve our onboarding efficiency, which will lower our CAC and allow us to grow more profitably. Then finally, on the fixed costs, you've seen over the last 6 months, after a year in 2025 of growth, we've been holding head count a lot more flat. In fact, in 2026, we've had a 4% reduction in the size of team but without any reduction in productivity or output because really each team member has been delivering more than they ever have. We're going to maintain a relatively controlled fixed cost base and allow those gross profits to give us more overall EBITDA minus CapEx. Then finally, sunsetting the legacy platforms. These are a huge when we sign we sunset Sanimalis in Q2. VetVision will be by the end of the year. And obviously, in our future, we have Aspit and Vetera. by clearing the way in these old platforms, we give ourselves the ability to focus on our core flagship platform without complexity without distraction, which saves cost and allows us to move more quickly. So we're -- we care deeply about getting all clients onto our flagship platforms. I also wanted to talk a little bit about the investments that we want to make during this 2027 to 2029 period. It's our intention to be a high-growth business for a very, very long time. And to do that, we are going to continue to reinvest some of the surplus profits that we're going to be making in 2027 to 2029. The first of those is we're going to increase our U.S. presence on the veterinary side. Local leadership local go-to market. We really want to increase our footprint in the U.S. in order to be able to take advantage of that ginormous opportunity that Charles presented earlier. And then on the therapy side, we have this huge addressable market that we can go after. We're going to need to invest in expanding to new markets and new product development. All within the boundary of making sure that on a group basis, we are cash flow positive and EBITDA minus CapEx positive. That's it, to reiterate the 3-year guidance, 15% CAGR, between 2027 to 2029, plus or minus 2%, driven by veterinary at 20%, plus or minus 3%; and therapy at 10% plus or minus 2%, all the while, we're going to be adjusted EBITDA minus CapEx positive and cash flow positive, excluding any potential M&A. We will not be sourcing external financing to fund organic growth. And with that, I will turn over to Charles to conclude the session.

Thank you very much, Alex. Thank you. So you've heard a lot from us today. So I just want to summarize the key takeaways. So one is across both business units, practice management software is the core infrastructure of the clinic, and it gives us a unique right to win to be able to build great agents on top, right, because we control the data, the workflow and so on. So we have the unique right to have a way better agent than others that are stand-alone. The second is that we own the #2 vet PMS globally and the #1 in Europe and the #1 therapy PMS in the Nordics. So we've got a really strong market position in the market that we're in, right? The third is the low churn, right? Customers stay with us for a long time and trust us with their business. And at scale, we can see very strong unit economics, not only on the -- the net upsell, right, which is a churn plus net retention, which you'll see from increasing the average revenue per location, but also in the sales efficiency that we've got in terms of recruiting new customers. And what that yields is the profit margins that we see in the Nordics for veterinary, which are around 40% EBITDA less CapEx or 45% for therapy. And we also have great headroom for growth. We've got a great opportunity to continue to expand. We're only a very small percentage of the veterinary market in our current markets, and we've got a great opportunity to conquer the rest of Nordics and go beyond therapy. So -- and on the guidance, Alex went through it, but we're looking forward to, as Alex mentioned, it's really important for me to make sure that we are never reliant on anyone else, right? And also for our partners, customers that we always have enough cash and some buffer to be able to be masters of our own destiny. So being cash flow positive and having a growth rate of 15% plus or minus 2%. Thank you very much for your time, everyone. And I think next will be Q&A.

Thanks a lot, everyone. Thanks, Charles, and to all the speakers today. We're going to take some Q&A now. So we're going to bring some stools up on stage, and we're going to invite the 4 speakers from Nordhealth up on stage. Q&A is open to the room. There are people with microphones who will hand you a microphone. [Operator Instructions] Great. [ Torbjorn ].

Just wondered if we could start with a comment on AmeriVet, the rollout that was canceled earlier this summer. Maybe touch upon what the reasons for the canceling of the rollout was.

Charles, maybe you want to take this one?

I'll take this one. So we originally had a CEO of AmeriVet, which wanted to build a product together with us in the U.S. That CEO used to work for VetFamily, which was part of -- is now part of IVC. And they -- that's what we did with IVC. So we're doing with many of our partners here is that we work together to localize the software for that market together. There was a CEO change, the management change and that strategy changed in that they wanted the software that was localized from day 1. Even though the technology is quite old, they decided to go for the current solution that's localized fully for the market.

And a follow-up to that one because given that AmeriVet was an important proof point for the U.S. market, has it made it more challenging to get into other processes with other potential customers in the U.S.?

We've got other corporates like PetVet365, for example, which rolled out in the U.S. and we're also talking to lot of other corporates. And there's a -- the market in the U.S. for corporates, there's 2 players. It's either ourselves or ezyVet for now. And yes, ezyVet have a broader array of integrations than we do currently. And -- but that's only a matter of time. So if -- once we do have that only remaining gap solved, right, I think we'll be in a very good position to be able to win all of them. And we've got PetVet365 is a good example of a good corporate for us in the U.S.

And I can just continue. On the 3 clients that you mentioned, the 100-plus clinics, clinic clients, what geographies did you did you sign those in? The 100-plus clinic clients that you mentioned, what geographies are those clients connected to?

We're not disclosing specifics on the clients. I mean, it's -- the number of enterprise clients you can see per geography is relatively low. And so -- at least of that kind of scale. And so while we're in pilot and early implementation, we're not yet disclosing the geographies or the specific client names.

The reason is, if I hear of a pilot, I will go after that customer because I know they're willing to switch PMS. So I just still want to provide that opportunity to potential competitors.

Understood. And on the upselling opportunity, particularly within veterinary, both payments seems to be a very large part of that. So what is really -- what is the key selling points for clinics to use your payment solution instead of a third party's one?

It's a very weird market in that, actually, this is not an opportunity for corporates that much -- or it is an opportunity that they can get a good payment solution, but pricing, we basically provide that cost to our partners here, for example, because they get very low rates. Smaller, independent clinics have very high rates relative to the enterprises. And so the value prop is you pay the same as you pay today, but you've got a fully integrated solution. And an integrated solution means you don't have to key in manually the amounts, so less errors, less work. And secondly, which is fun for the accountants here, is that invoice to payment reconciliation is fully automated. So you have to spend way more -- less time to be able to do that and way less time if you have a third-party accountant to be able to do your books. So the opportunities on the noncorporate side, at least on the revenue opportunities, right? We still provide that to corporates because we think it's a great service, but that's more -- we don't make a significant amount of money from that.

And how -- on the unit economics side, you touched a little bit on it, but both for payments, but also for the AI solutions that you are searching to upsell to those customers. Yes, how does the unit economics...

So when we look at the unit economics from the slide, actually, that's actually gross profit. So it's the way that accounting works for payments is that we take out all the costs from -- which are normally made up of 2 things. One is interchange, which is the amount of Visa, Mastercard charges. And then the second cost is scheme fees and also the added payment fees. So this is just on top of that, what we make, right? We have cost in terms of like providing support and for that service on our side. Then on the AI side, the cost is mostly on the development, right? It's a -- yes, the cost of AI models is increasing if you've got the latest tools. But like the -- if you want to use the latest frontier models for everything, it's also very slow. So actually, it's not always the best use case. And so who knows where model costs will end. But currently, I think that the gross margin of those is not the real issue. The real issue is that developing it and iterating on it, that's the main cost. So it's a fixed cost versus a unit gross margin issue. Gross margin is probably similar to the software.

[indiscernible] to that as well. I mean, in terms of AI models, we're not reliant on any particular provider for any kind of special functionality. And all of the AI that we've done in the product is built in such a way that we can very easily switch to different providers if we need to. And I think more recently, it's been quite interesting in so far that there's been a huge amount of Chinese open weight models that are very cheap. And if we wanted to switch to them tomorrow, it's like a few lines of code and we've switched and the kinds of things that we're doing in the products, we're not reliant on any particular provider. So I'm not worried about cost of using AI at all because I think the market itself is keeping prices down, and we're completely agnostic in terms of our solution.

And cheap also means faster. So actually, it's not -- it doesn't mean that if it's cheaper, it doesn't mean it's less good, actually. Sometimes there is cheaper and better because it's -- the processing is much faster.

Yes. And I think the other thing that's interesting as well is that we don't have to use the latest frontier model that comes out. Up until recently, we were using models from 9 months ago. They were amazing 9 months ago. They're still amazing today. So whatever the sort of narrative out there of the cost of AI, we have complete control of what we use.