Welcome to Carasent Q4 report for 2025. For the first part of the conference call, the participants will be in listen-only mode. During the questions and answer session, participants are able to ask questions by dialing pound key five on their telephone keypad. Now, I will hand the conference over to CEO Daniel Öhman and CFO Svein Martin Bjørnstad. Please go ahead.
Good morning, and welcome to our presentation of the fourth quarter, 2025. I will start with a business update, and today a little bit extra about AI again, and then Svein Martin will give a financial update. As most of you know, we're providing EHR systems for healthcare providers. We're growing quite rapidly at 15%. We have more than 90% recurring revenues and a net revenue retention of 110%. Looking a bit at the quarter, we have a strong quarter with high activity, especially given that we have taken almost all of Volvat, so that's Capio Norway online. That's something we're really proud of. It's more than 5 million medical records have been transferred from multiple old systems into our system. So I think, it's gone more or less on time and in a really good way.
So we're really happy with that, and it gives us an extra strong quarter with some extra high consultancy revenue. Even if that's not our focus and we price it lowly, it's the result of the Volvat implementation, mostly. In general, we have a really good improvements in profitability with EBITDA margin of 16% in the quarter. That's up from 5% last year, and there are no adjustments this quarter at all. We also have a very promising sales pipeline. We've been doing quite a lot to strengthen the sales team and marketing team in Sweden, adding new roles, and start working in more, in different ways. So really promising start, this year and the end of last year on sales, so we're happy with that.
Looking a bit into 2026, I think it's important to note that the Nordics, we're happy about. We believe that will grow strongly. We're happy about Germany, too. It's according to plan, but the old products in Germany are not growing very much, and the new product will start to roll out this year, but from a very small base, obviously. So that means that the German growth this year will be quite low. The Nordics should be good, and then over time, Germany will increase, giving us the three-year targets. If you're looking at growth in the quarter, signed not implement ARR amounts now to SEK 5 million. That's because we implemented Volvat, most of it, and also Frelsesarmeen in Norway. But we have a strong sales pipeline, so we're not worried about that.
In total, we have 15% organic ARR growth and 16% reported organic growth. If we look at the entirety of last year, we're really proud and happy with what we have accomplished, with strong improvements from last year. We have quite strong growth, and you see the EBITDAC, so that's our main target that we look at internally, EBITDA minus CapEx, going from SEK - 30 million to SEK +34 million. This is something we aim to continue with, having a very large percentage of our growth falling down to higher margins and especially EBITDAC margin, which we think is the most valid number to look at when it comes to us. What's almost everyone is talking about at the moment, AI, and we will be looking at three different perspectives this morning.
So what does it mean for our development? How do we use it in product, and how do we think about it? And what about our position in the market? So if you look at development or productivity, our strategy has been the last couple of years since I joined, to have a high development pace and to develop our products faster than the competitors. So by always having a higher pace than our competitors, and our competitors are old legacy softwares, the distance between them and us increase every year, and thus increasing the need to change system and make it easy for us to convince customers to change to a better, newer system from us. With help of AI, we can develop faster, obviously.
That means that compared to the legacy softwares, we will improve more rapidly, and the change will become bigger, faster. We continue on that, in that plan. I think it's important also to use the productivity gains to improve the products faster than before, because what will happen when productivity and development increases is obviously that development will go faster, and by keeping a high pace ourselves, we stay clear of new entrants, and we gain on incumbents and old softwares. We'll continue that. But having said that, it's important to, at the same time, make sure that we always fulfill all legal requirements. There are strong, high demands on how we are allowed to develop and in what way, and what process we need to have, from, most of all, Läkemedelsverket in Sweden.
But there are similar rules in other countries, and we cannot and never have disruptions to our services. If we have, our customers cannot work. So the most important for our customers is that software is always up and running, and that's 24/7. So we have to maintain that same time. But having said that, I think that the new tools are great. We're using them more and more, and really getting productivity gains from them. When we look then at how we use it in the product, our plan is to continuously add AI functionality in the product. I think that most of you know that a year ago, we decided to increase our spending on AI in the product, as we saw that it can really help our customers become more productive.
So we aim to add more and more functionality in the product step by step and giving it more of an AI agent feel all overall. The first end product is Medsom, which is an ambient scribe. That means that we listen in on the call, the discussion between the patient and the doctor or the physiotherapist, and then we automatically propose a medical note from that. That's the functionality that's gotten the most use within healthcare so far, but even if this is the most used type of healthcare, so productive function in healthcare from AI, it's still way below 10% of our customers that use any of these type of tools. So it's not moving very rapidly, but still, it has the potential to really save time for our users, and that's something we're working on.
Given that our customers are afraid of change and especially our strong position, we don't need to be first. We think it's good that our others are first, and that we can see what works and doesn't work, and in what way our users like functionality. And if we believe that we can do it well, and if it makes sense to have it in our systems, then we can develop it. So we don't need to be first, but we should really follow very closely after. So taking Medsom, then ambient listening, where there are a couple of products like that out in the market already, still, use is very, very low among our customers, but we believe it will grow over time. So then we really need to get there.
Given our product, then Medsom, we have been a bit too slow, I think, even given that we don't want to be first, we should be a little bit closer to being first. The reason why we've not been in the forefront is that we have tried to stay clear of U.S.-owned clouds. Even if they are in Europe, we don't want to be in them. It's a promise we make to our customers because most of our customers do not like to be in the U.S.-owned clouds, and many of them know that it's quite likely that it will be illegal to be in U.S.-owned clouds with personal data in the EU again. It has been twice before, and our customers are really, PR, because they, the type of data they are handling, and that's why they are very, very well informed about these type of risks.
But what we have noticed is that the type of a LLM models we can run outside of U.S.-owned clouds are, they are too weak, they are too small, so it's been Llama we've been running, and it's not good enough. So now we have moved to OpenAI and ChatGPT, instead for our customers, and we'll be rolling out that. Those who really don't want to be in the U.S. clouds, even if they are in Europe, they can still use Llama, but I don't think that our customers will use that. But we can always switch them back to, for example, Llama, if we want to, and they want to. And now we're really getting the pace up.
So last of all, the market position. So EHR systems, by nature, I mean, they are truly mission-critical. Our users, the healthcare providers, they are standing still when our systems are not up and running. They don't know how to meet, who, which patient to meet. They don't know anything about the patient. They cannot make sure they are called, they cannot charge payments, they cannot do basically anything. So, and on top of that, we have countless integrations with different levels of Swedish society, and different types of entities. There are very many different workflows that we need to accommodate in our systems. Very many different people that collaborate within a hospital or a healthcare clinic. There are tough legal requirements, and it's really painful to swap systems, or that's what our customers think.
We're trying to change that as much as we can, but our sector, they don't like change. Doctors, nurses, and so on, they are trained never to make mistakes. They don't like taking risks, and changing systems is a risk as they see it, so they avoid it. And also, EHR systems cost only 1%-2% of the turnover, so it's never worth taking any type of risk for them for that cost. So that's always our challenge. What we are competing with mostly are systems that are 20-30 years old, and they look like they are 20-30 years old. We have a lot of new good functionality, and still it's hard to make them change. So in total, I think that there are very few software or knowledge companies that are, have a stronger position than us in this environment.
I think we really can gain advantages for this. We can add more functionality to our customers, we can help them help more patients. We can also improve our products faster than before, and since we are the ones investing most in our markets, or most of our markets in our products, we are likely to gain the most, I think. But having said that, I think it's always important to stay curious, stay on our toes, always think about our position, how can we strengthen it? And that's what we do every day, and that's what we've been doing for a long time. But I think we will. This is a discussion that's likely to continue. But for now, I will continue to the next slide and look a little bit at the journey ahead.
So, we aim to continue with the high growth we have. We have added sales resources in Sweden, also, divided workflows a bit differently to make sure we have a strong focus both on existing customers and new customers, and also working more with outreach and sales. We will now strengthen the sales and marketing staff and organization in Germany in order to roll out the new system. And it's a really exciting time, and we have many projects that are now seeing the light of day and are ready for commercial rollout or are in commercial rollout. So we have WebCare in Germany, we have our new system, Adopus Web in Norway, we have surgery functionality that more and more customers are adopting in Sweden. We have Medsom for ambient listening and so on.
So, we're really happy about where we are and find it this will be a very interesting and fun year. And as I mentioned before, we aim to continue to invest a lot in our products, but we do not plan to invest more than we do. So, we aim to every day get more efficient, being able to get more functionality to our users with the same resources, and to help more users, and to sell to more users with roughly the same resources. So every day, getting a little bit better, and over time, that adds up to a lot. And we maintain good cost control.
What will be really fun and interesting this year is, of course, the launch of Webdoc X, where both our team in Germany, the early pilots, and when we show it to on fairs and so on, we have really strong and good reception. It would be a very interesting year to follow that and see the first customers, how do they feel about it, and then start getting sales going. With those words, I will hand over to Svein Martin.
Thank you, Daniel. Looking at the highlights of the quarter, Q4 was a very strong financial quarter, where the profitability improved a lot, and we also had good growth. ARR ended at SEK 331 million, including the signed contract, and we grew our ARR 15% organically with 110% net retention, and the margins on EBITDA was 25% and EBITDAC 16%. So starting by breaking down the growth, one point here on the high level is that it's the first time we include the German acquisition in the organic growth, given that it's been part of the group for over a year now. And that is still all the legacy of products, so it's not growing, basically.
So, the Nordics basically means that the Nordics grew very well to be able to maintain the 15% for the group in Q4. Looking at the net upsell, important contributor here is Volvat. It's included in net upsell because we had the very small part of Volvat before the new contract. So, that contributes with NOK 6 million. And, worth mentioning as well is that, new clinics for existing customers, groups, such as Capio, is included in this figure. So net upsell has been on the sales side, very strong in 2025. The churn we have spoken about many times over the last year, it's elevated figures, mainly due to non-voluntary churn, such as bankruptcies, that we saw last year.
On the churn, we generally are notified by our customers canceling the contracts at least three to six months before it takes effect in the figures. What is good is that we had, in terms of like canceled contracts before it's reflected, we had high levels in second half of 2024 and also in first half of 2025. But for the second half of 2025, we saw that the churn levels dropped again down to normalized levels, which is positive and will be reflected in the figures basically this year. New customers, it looks like a small percentage of the revenue growth is 5%.
But what is positive is that if we compare it to what the same figure was last year, it was SEK 9 million, so it actually grew 56% year-over-year. So it shows that we are succeeding well with our sales. And this is, of course, also very important driver of net retention after the customers are onboarded, as they grow with us. Moving on, looking at the P&L, our revenues ended at SEK 93.6 million, compared to SEK 78.7 million last year, so we grew 19% in total. The growth was boosted by a very high consulting and other revenues. If we look at the gross margin, it was impacted that we have this one big one-time revenue in Germany that had direct costs linked to it.
So this increased the COGS. And if we hadn't had that revenue, the gross margin would basically be stable year-over-year. Also looking at the cost base, it's developing according to plan. We did have a positive effect by a grant in Norway related to one of our development projects, where it basically reduced the costs in the quarter by around NOK 2 million. So this was according to plan, and we had expected it, but it's nothing that's going to come every quarter, basically. So this resulted in a big improvement in profitability. So you see here that the EBITDA increased from 12 million to 24 million, on an adjusted basis last year.
No adjustment this year, and the EBITDA landed at SEK 14.5%, compared to SEK -16.8%. But on adjusted level, we increased from 5% margin to 16%, which is a big improvement. Naturally, it's worth mentioning as well that this quarter is a bit special in terms of profitability, given that we have the high consulting revenues and combined with the lower cost that I mentioned. So if we hadn't had this, the margin would naturally be a couple of percentage points lower, like our underlying run rate. But still, very good improvements year-over-year. This slide is a good illustration of how we think about running the business. Basically, it shows our revenue development compared to our costs.
The orange line here is the revenue LTM, going from SEK 245 million to SEK 344 million, growing about SEK 100 million. We have the gray part of the bars is the COGS, which increases a bit, and we have the light blue, which is the acquisition in Germany. That was done in Q4 2024. The dark blue part is the cash cost base of the organic business, basically. OpEx, CapEx, personnel expense, and you see that it has been very much stable over the last couple of years, which means that the business is scaling nicely.
For 2025 in total, we were able to convert around 80% of the revenues into profits, which is basically in line with our target, and we still invest a lot in our products, so we aim to be able to scale the business for a long time when we grow our revenues. The key is to grow the revenues and keep the costs under control. Finally, on the cash flow. Cash flow is naturally supported by the big improvements in profitability. However, we see that the working capital development in 2025 has been very weak, and there is a few reasons for this.
It's mainly these three reasons that I list here. Firstly, as I've talked about many times during last year, is that we took, t he relisting costs were taken in December 2024, and then it was paid in Q1. So that hurt, you see the effect there in Q4. So it had a big positive effect on the working capital in Q4 2024, and then a big negative in Q1 this year, 2025. It's around NOK 15 million effect. Secondly, we had a large customer that, due to a mistake on their end, didn't pay their invoice in time in December. This is a big public healthcare provider, so it's a very solid counterpart, but either way, it was paid in January this year instead of in Q4.
Thirdly, these big consulting revenues that we have talked about mainly occurred in December with the payment date in January 2026 as well. So it didn't come in Q4. So as you basically understand, some of these effects will have a positive impact on Q1 2026. Then we have done the share buyback program, which also impacts the cash position. We have in total acquired almost 6% of the outstanding shares in Carasent. But in summary, another strong quarter financially, very much in line with the plan that we have. And with that, we can open up for Q&A.
If you wish to ask a question, please dial pound key five on your telephone keypad to enter the queue. If you wish to withdraw your question, please dial pound key six on your telephone keypad. The next question comes from Fredrik Nilsson from Redeye. Please go ahead.
Thank you. Good morning, Daniel and Svein Martin. I want to start with the situation in Västra Götalands regionen. I mean, they have more or less decided to go for a modular solution. Have you seen any impact on your ability to approach primary caregivers in the region so far?
Good morning, Fredrik. So, yes, we're targeting primary caregivers in Västra Götaland region all the time at the moment. They are never fast, but I think we have good and productive discussions with the potential customers, and we have gained a few. I think that if you talk about the situation more in large, we are also hoping that potentially the public primary care could come into question if they're going for as they are planning to go for modular approach, but that will take time. The last two attempts to implement softwares, so Millennium, they never succeeded with, but also the one before, it took eight years. So, we assume that the situation will be more or less like this for seven to eight years from now.
So, a nd I think the caregivers think the same. So, now they're open for discussion, and we have good productive discussions, I think.
Okay, great. And also, I want to continue a bit on your discussion about AI. You mentioned that you have changed Llama to OpenAI. When did you do that, and is there any impact on gross margins, pricing, and what's the feedback so far?
Yeah. So it took quite some time to change actually model, 'cause we need to have guaranteed the clouds in Europe, and we also need to have zero retention. So we only have the data in the Azure cloud for, I mean, basically, yeah, almost no time at all, and then it's completely gone. OpenAI, I think they have some growth pains, maybe. It took a really long time to get an agreement in place. It would have been much faster with Google or Claude, but when we did the blind tests, we saw that OpenAI was usually the one doing the best ones. So we got live last week, and our pilots, we have a number of pilots running in Medsom, and there are many of them also running competing products.
So, the early feedback is that we are on par now, 'cause we've done a lot of other changes, too, to that product. So we feel we're in a good position, and we will want to see one week more of feedback, but then I think that we will just push sales on Medsom now. We have been pausing sales in Medsom for a while until we felt we had the same quality. We don't want to take customers online, and then they move back, 'cause then it's really hard to move them back to our product. So we have taken it a bit easy, but now we'll start pushing sales again on Medsom.
Great. That, that's helpful. And you talk about strengthening your sales organizations in Sweden and Germany. What magnitude of increases are we talking about approximately?
So, as I think most of you would know, we don't have a very big sales organization. But basically, for our main product, Webdoc, we have gone from five or we had also people on sick leave and so last year. So last year, we had roughly four people in place, and now we are at seven. And what we've done is that we introduced customer success managers that are responsible for most of upsell, and then, the sales team, their role is to work on leads and also, reach out to new potential customers. And, we also added some people which we believe will really help us sell. And it seems to be working well. I must say, I was a bit, i t's been working better than I thought so far.
To really convince someone to change systems, that's quite hard. But most of our customers, they feel that they have a poor solution today, the potential customers, and therefore they are open to discussions. So I think the sales teams are doing a great job, and I think this change have really worked. But step by step, we're getting better and better. That's the way we've gone from roughly four people to seven people in the Swedish organization.
Okay, I see. And last question from me. I mean, regarding the surgery module, are you seeing any meaningful contributions to ARR near term, or what's the status of that product?
The status of the product is that the development of the product, so nothing is ever completely done, right? But we will wind down the development of that product in a couple of months; then we feel that it's really finished. We think, we felt it's been good enough to sell for like half a year or a little bit more maybe even, but there's still been functionality that's been important for our customers that they've been wanting. So a couple of customers, they've been wanting certain things, and we've had them live, but they will start paying them. But the main importance for the surgery module is to win new customers and large new customers. So on average, the customers that need that module have maybe 150 employees, somewhere around there.
We have a number of those type of discussions, and that would really, you know, help our new sales, 'cause most new sales today are startups. So that's why even if we add a lot of customers today, in Svein Martin's picture earlier in the presentation, it's still quite a small part of it, and that's because they are new opens, and then we grow with them. So that's great. But if we can add some big customers into that mix, it really helps our growth, such as Volvat, for example. So we hope and aim to win a number of large clinics this year with the help of the surgery module, and I think that it will help us over time to win more and more of that type of customers.
Great. Thank you very much. That's all for me.
Thank you, Fredrik.
The next question comes from Albin Sandberg, from SB1 Markets. Please go ahead.
Good morning, Daniel and Svein Martin. I hope you are well. I have a couple of questions here from my end as well. Perhaps beginning a bit more high level, it would be interesting to hear your thoughts about how one should view and kind of AI-related risks within your vertical, given the debate we've seen across the Atlantic now in recent weeks. That has also affected the kind of software assets here in the Nordics as well. So it would be interesting to hear about your thoughts about the kind of entrenchment and barriers and risks and so on.
Yeah, so, it's a bit difficult. It's a bit like when I was CEO of GHP and we had a government that wanted to outlaw private healthcare or insurance payments, and it's so many different scenarios that everyone get in their head before it's really concrete. So that's why it's difficult to meet everyone's view on this in one answer. But in general, I would say that, as I mentioned in the presentation, was that our position is extremely strong. And what we are competing with and our challenge every day is to get people to change from systems that are 30-20 years old, and they look like they are 20, 30, 20 years old.
We're trying to get them to start using functionalities such as self-checking for payments for patients so that they can save time. We're trying to get them to start having automatic calling for patients. So we can automatically send appointments to patients, and then the patient can change the appointment themselves, so we can save a lot of administration time for our customers. But they are really afraid of change, and there are a lot of legal requirements on how we're allowed to do it, and there are so many e ach time we develop a function such as that, for example, we have to make sure, take the self-check-in that, yeah, if it's an insurance patient, then we have to inform the insurance company, and then the patient should pay a little bit themselves.
If it's a patient from maybe Västra Götaland, they should pay another amount themselves, but then we have to check with Västra Götaland or with another region, depending on which region the patient is from. If how much if they are entitled to not tend to pay anything at all, and so on and so on. So it's, there are so many interactions in the software and so many different workflows, and we're also a very small part of the customer cost, but we're an extremely important part of them running their clinic. So we're talking about organizations that are afraid of IT.
They don't really have m ost of our customers are in the size of maybe 30, 40 employees. They hardly have an IT person, and with us, they don't have to have one. So I have a really hard time seeing that they should do something themselves or would do something themselves. I think that as long as we press on with development, we can become more and more efficient, delivering more and more functionality to our users. I think the software companies that are really at risk are the ones that have been milking positions. So I think that many around us, and I think that, I mean, a lot of investors have been telling us for quite some time that you have a really big cost for R&D compared to competitors and other software companies.
And the reason is that within our industry and similar industries, it's been possible to not develop your products and have really, really good cash flow. And as we've said the many times during the last couple of years, we don't believe in that, in that strategy. We believe in improving our products, so we gain more customers. 'Cause the milking position, and especially with AI, becomes weaker, much more rapidly now. So those, and those are the type of companies that we compete with, to a large extent. And that means that we will now gain on the, the advantage we have will become faster, much faster. And also for them to start kind of develop their products and start using these tools is quite difficult.
If you don't have a tradition of doing new development, you are mostly just adding more interactions, you're adding new legal requirements, but you're not really moving your product forward. 'Cause you've been able to feel safe in this industry by just the number of different workflows and all the different integrations you have done. And you've been keeping adding integrations, so over time, you have so many integrations and that's why it's difficult for us also to replace others. 'Cause we've been adding in integrations also for a long time. But now I think it's important to also improve your product, and that's what we've been doing, and that's what we continue to do, and we can do it at a higher pace.
So I think, and as I mentioned, also, I think that I have a hard time seeing any software company or knowledge company that have a stronger position than us. But I think also it's important to stay curious, to think all the time, how can we strengthen the position? How do we use this functionality to the best extent? How do we get it into the product? Already today, healthcare workers, so doctors, nurses, and so on, they go from using their iPhone or using modern software solutions in the private setting, and they see something from the 1990s, and it looks like it's from the 1990s.
I think it's important that with AI agents, likely on the rise, that also our systems feel similar to what they use in their, in their private life. So to incorporate that type of functionality in our products, I think is also important. But it's us who have the data. We're not allowed to give the data to anyone else. They cannot use an agent on it. It's, y eah, I think, I think we have a strong position, and then there are many different scenarios to discuss. Did that answer, answer your question, yeah?
Yes, it was a very extensive answer as well, but thank you, Daniel. Continuing a bit on AI, I guess, I mean, with the Medsom, which you've kind of gone with a soft launch with now, has that helped anything on the revenue side, or has it still been kind of beta pilots, still before a full-on commercial launch here, hopefully in the coming weeks? And furthermore, on the cost side, you mentioned a couple of quarters ago, the increase in costs from AI development. Is that similar? Has that been kind of a similar cost pattern now, and do you expect kind of the same going forward, or will that an increase, or will we perhaps see a net decrease, given that you can become more effective and so on?
So, if you're talking development in general, I don't think it's the time to invest less in development. We might have it in different patterns and in different ways than traditionally. But I think that these productivity gains, just as we had when we had factories coming for the building of cars and things like this, we have better cars, more cars, more functionality in the cars. I think it's the same for software. We have to move faster. So, it will change a bit how we work, and it is changing our work, but I don't think it will, I don't think we should decrease costs. I mean, we can do the same pipeline that we had planned on the roadmap, and have less cost, but I think that we should put in more in our roadmap.
When it comes to Medsom and costs, we will add a bit more, what's the word? Staff to it, we can say. But it's within our targets. It will not change anything drastically, but we will add a bit more and focus a bit more on it than we have. But we already have quite a team working on it. It's very easy with the AI to get it 90% of the way, but we need to be 100% of the way, and then it always gets, i t's the last 10% that are tricky and difficult.
So, that's why we will continue to invest in the AI in our products, so for our users, and I think if anything, we will increase it, but then we will also think about costs in other places in our structure and move more focus on this part. So as always, we have so much we can do for our customers to help them help more patients, and we always prioritize what we think can help them the most and where we can gain the most value, either through that we can win more new customers more rapidly, or we can increase upsell. And AI agents in the products is definitely an area which can help our customers save a lot of time.
The main burden for our customers are, is administrative. So they have a lot of administrative tasks, and this is where AI is good, but it also has to be near perfect, because they are liable to what's written, for example, in the medical records. It has to be fully correct. The customer is looking at it, and it, I mean, it can be that the patient is dependent on certain medication. It has to be correct. It can be that they are allergic to a certain medication, that has to be fully correct. It can never be wrong, and the doctor is themselves liable if it's wrong.
So then that's why they are slow at adopting these type of things, and that's why we have to really make sure it always works really well, and at the same time, it fulfills all the legal requirements and so on. So it's, b ut we will invest more in that, but I think it will be on the expense of maybe having other things in the roadmap, but it also will come from us being able to put more in the roadmap in total with the same cost.
It's also worth mentioning, in the short term for Medsom, when we change from our own hosting to OpenAI, it also means that we had to sign up for some minimum volume. So, the cost for hosting will increase. When we have fewer users, it will increase by around SEK 1.5 million compared to where we're at on an annual basis. But then when we have many users, the gross margin will be better because we get a lot of tokens for, for, that minimum volumes.
Okay, great. Thank you. That's very clear. And when we kind of summarize, I guess, the increase of sales and marketing in Germany here in 2026, and also these kind of, how to say, more projects with putting into the product, how should we think of kind of overall on the OpEx side here in 2026? Will we see a kind of similar organic OpEx increase as we saw in 2025, or is there a reason to believe that you're kind of accelerating this now? How should we think about that?
So, the cost where we will add a bit of cost in 2026 is in the German sales and marketing team. So, that there, we plan to add a few roles, but in the Nordics, we plan to maintain a similar staff level. We add a few people on sales in Webdoc, that's pretty much it.
You will not see any drastic changes, and we always, so say, we're always adopting organization. So even if you didn't see too many changes during last year, for example, in the total cost, we did a lot of changes in the organization structure with people that had to leave and new people coming on board. So we always try to have the optimum, so to speak, organization. So we continue to make changes, but, but the overall guidelines is where we aim to continue, and, and, a large part of our growth should fall down to the bottom line, and, we, yeah, we continue that way.
Yeah. I think a fair way to look at it is basically a similar cost base, then we also, of course, have like wage increases and inflation and a handful of roles, mainly in the German business.
Okay. That's also very clear. Thank you so much. I think that was all from my end right now. So thank you much for taking my questions, and I'll get back in the end of the queue here.
Thanks, Albin.
Thank you.
As a reminder, if you wish to ask a question, please dial pound key five on your telephone keypad. There are no more phone questions at this time, so I hand the conference back to the speakers for any written questions or closing comments.
So we have gotten quite a lot of questions in the chat as well. We might have addressed some of them, but we can start going through it. First one from Rickard, I can take this one. Was full recurring revenue from Volvat live during all of Q4? No, not during all of Q4, but in the ARR, it's basically implemented. But it took effect from the mid of November in Q4. That was when the big system was migrated. Next one from Risto, h ow do you envision the profitability of those AI functionality tools from a gross margin perspective?
Yeah. The gross margin is a bit lower on those type of tools, as we have the cost for either OpenAI or Google or Claude. But it will not be dramatic, as we, in our targets, still don't believe that we will have dramatic revenues from those type of functionality in the near term. That could be that it's will grow faster than we thought, but our industry is slow at adopting new technology, ambient listening. So when we automatically propose medical notes, is something that really can save time for each visit. It's been very widely discussed now for two years in this sector. It's there has been good solutions for it in at least one and a half, two years. But still, the adoption is really low.
So, but what we see in the U.S. when it comes to ambient listening, U.S. has much less regulation and a much more commercial healthcare, so bigger pressure to become efficient. And the U.S. healthcare works very differently from the European ones. But there you have much more ambient listening and for just standalone ambient listening products, there have been a price war. So, if you're just like a ChatGPT wrapper, then of course it's difficult to maintain high margins. I think for us it's more of it's part of the system as a whole, and we have a lot of information that we can use, so we're not really in the same type of competition, I would say, not over time at least.
But it depends, if we get a much higher and quicker uptake than we think, then we'll grow faster than we think, but then our costs will also. Our margins on that additional growth will be a little bit lower than we're used to.
Next one. Given your current stock price, does this increase your appetite to increase the capacity of the share purchase program?
Yeah. No, we're trying to maintain a steady pace. I think it's. And it's basically up to the board. It's not for me and Svein Martin to answer. I would say that as CEO, we should never really have a view of what's the right share price for your own company. It's up to everyone else to judge how we're doing. So, the board will decide on that pace. We have no new information to give on that.
Yeah. As we generate more and more healthy cash flow, we have found that this combined with M&A is a good way to distribute capital, basically. It's a long-term effort, and in addition to the one we have this year. Next one from Niklas, y our strong balance sheet and accelerating profitability means that you have a lot of room for acquisitions. Have you seen any changes in the M&A market during the current turbulence?
So we are looking at a couple acquisitions, as I think most of you know. It's mostly acquisition that strengthen our position in the Nordics that we're looking at. Where we have really, yeah, clear synergies in, in getting a stronger position, for our present products, mainly Webdoc. So we're in those, we work on that. When you have a very, you know, narrow target focus, and the ones you're talking to are probably not for sale, so you're approaching companies that you think are interesting and would strengthen your position, we have maybe, like, five companies on our list, then it always takes a lot of time. 'Cause they're not structured processes, the, the owners might not be willing to sell, they have not planned to sell at least, so you have to convince them.
But my experience is that then you, that's when you do the really good acquisitions, 'cause you can target ones that really, you really believe in and that you really believe can add value to, to what you do. When it comes to valuations, I think that the private market is always much slower than the public market to changes. Typically, I think that even when the public market has gone up or down, the private market is quick on the up but really slow on the down, because if you're on the private market and you don't have to sell, you just wait in a period like this.
Great. Next one from, Rickard, a re you following the original timeline in Germany, as far as pilots, commercial launch, and sales, expectations?
So, not fully, I would say. I would have liked to have been a bit further with the product at this point in time. I think this is quite often the case, that when you start developing, and even when you have a really clear answer on what to develop, once you get into the real details, so how is this supposed to be reported to a certain authority in Germany, for example, a certain type of information? And when you dig in really into the details, it typically gets a little bit more complicated than you thought from the beginning. And that's the same for each one of all those type of things. So I would have liked to have been a little bit further, but I still think that we are in a good place.
We have a really high pace in development, and the product looks great, but we are a little bit behind the original timeline.
Next one from Niklas, w hile your platform is mission-critical and difficult to replace, there is a lot of third-party integrations to your platform that are sometimes easier to replace. Are you looking to increasingly develop your own solutions to replace third-party products now, when it's easier to code with the AI?
So, not from that sense, but I think that with the help of AI, our products can target a little bit wider group of potential customers. So, what's difficult with software, if you want to have it without an end date. So that's what we've been talking about all the time internally. We don't want to have an end date of our products. That's why we've developed a lot, we've put a lot of effort into our products. That means that you always need to make sure that all parts of your products are more or less up-to-date. So the first parts you wrote, the code, first code you wrote has to be up to date also 10 years later. So we need to replace the code all the time, so the maintenance gets quite high.
At the same time, you want to build functionality for new customers. That means that you, over time, have a very, a lot to develop each year. So, that's why, for our products, we have said that this is the types of customers that this product is aimed for. With the help of AI, we can be quick in development, we can get better at maintenance, so possibly we could have a bit wider scope for our products, when it comes to different types of customers and different types of customer scenarios. So that's how I mostly think about it. It could be a little bit that the integrations, but many of the integrations we do are to, like, insurance companies, public entities, quality registries, and so on.
The other types we do are typically very specific functionality, and there, possibly, we can move a bit in that direction, too.
Okay, next one from Mark. Could you please elaborate on agentic AI workflows with regards to your product pipeline and R&D? I'd also be interested in your pricing model and your thoughts around the idea that seat-based pricing might negatively be impacted going forward, if relevant. Thanks for the good results delivery.
So starting with this last part, in general, we charge our customers for the value we create. So we're trying to have the same models that our customer get paid for. So in primary care, they get paid for each listed patient they have, so we charge them for the number of listed patients they have. In specialist care, they typically get paid for each patient they see, each visit, so we charge them per visit. So that's how we have been doing it, and that's how we plan to continue to do it. There are other scenarios too, but that's our, that's how we usually work.
When talking about the agentic AI workflows, yeah, we are planning to add more and more functionality, as I spoke about earlier, in it, and I think it's important that, when you meet our softwares at the workplace, at the clinic or at the hospital, it shouldn't, it should be similar to the type of softwares that you use in private, and today it's not, for our competitors. So typically, if you're a healthcare worker, you come into work, and you meet a software that looks like it's from 30 years old, and in the private, you're using an iPhone. So, as AI agents get more and more popular, we have to have similar experience in, in our systems. And I think we can, we can move, close to that.
There are a lot of legal requirements, and there are also, i t can never be wrong, as I mentioned before. It always have to be correct, which is a bit of the challenge. And we also have to make sure that we never expose any personal information in the wrong way. So, there are a lot of challenges in it, but we're working on it. I think that we can add really good functionality step by step, helping our customers save a lot of time. So, it is an increasing focus for us, and it's been for quite some time.
Next one from Rasmus. Could you clarify on the OpenAI LLM? Is it run on their servers instead of yours, as Llama is?
Yeah. So you cannot run Claude or ChatGPT or Gemini free on your own servers. There are LLM models that you can run on your own servers, such as Llama or there are some Chinese ones and so on, but they are typically, w hat we notice is that they are not good enough. So now, the LLM work is done in the Azure cloud in Germany, but the data is there very momentarily. So that means that instead of having a high fixed cost, which we had for our own servers, 'cause the servers are really expensive when you have a lot of NVIDIA chips, we are using that cloud for just that functionality.
But as Svein Martin mentioned also, is that to have zero retention and have promise of to have it in Europe, we have to have a minimum use of tokens. So the fixed cost is actually a bit higher at the moment. But over time, it will not be a fixed cost, so it's a bit different. Then it can be that for certain functionality, we can use our own servers, and we can use smaller models and so on. I think it's good that we are working with both and try to find the right way of handling each type of data at this point in time.
Great. Next question from Yannik. Could you comment on the next steps you will take in Germany? Is the product ready to be pushed already? How will you approach sales, and will you need to invest further for that?
Yeah, so we'll add some sales people and marketing people in Germany. We'll also modernize a bit how we work in Germany. We are selling the product, so we have contracts for new customers, but we have no paying customer using it yet. So that's where the stage is. I think we should have paying customers in roughly halfway through the year, somewhere around there. It will not be a substantial impact on the revenues of Carasent this year. It will not be meaningful from that perspective, but it will be very important to start getting new customers. Sales of a new system, in my mind, is exponentially.
So you start with a few customers, you get a good reputation, you fix the first kind of things that they think are not really like they should, it should be, and then you start adding more and more customers. So it's really important to get new customers into product, and also get paying customers. 'Cause pilots is one thing, they are not paying, so they are typically not as critical to you as a paying customer.
Next one, Daniel, could you give a bit of more comfort that your pipeline is on top of the latest cloud product offering from a few weeks ago? Your thought on this would be much appreciated.
Yeah. So, it's a bit difficult to answer it completely. We see, as I mentioned before, very low risk that any of our customers would do anything with that product, and it's not really how the systems are used. It's really complex flows with many different user use case, different integrations, and so on. Our customers are not, I mean, we have a hard time explaining to them what the cloud solution is and the advantages of it, so they're very far from it.
And also, how you store data, in what way, and you have to make sure that each data is only accessible to the right person, and we have to store exactly which person to access which data, at which point in time, and the legal requirements of our customers, that they should take random looks at which one looks at which data. It's, it's, there are so many different requirements, so even if coding was completely free, there's still so many requirements and complex things and legal requirements on it, that it's very hard, extremely hard for someone to build a competing product.
So, but I think, as I said before, I think it's important for us to maintain a high velocity of improvements and always moving forward, using technology as soon as we can without breaking laws and without risking that the system goes down at any point in time. So as I said, I think there are very few with a better position than us. And then I think each one has to think, what does the future look like? But I think there are very few with a stronger position than us.
Next one from Rasmus. Could you talk about where you will switch the focus of development when the surgery module is done?
Yeah. There are some integrations that we want to do, and when I say integration, it's not just transferring data somewhere. We have to also create some new views and so on. That I can talk about later on when it's done. But that's what we want to prioritize, because it would really help some of our customers. So we can talk about that probably in the Q2 report. We want to talk about it when it's done and it's out. We don't want to talk about it before it's done. So that's, but it's something that we really look forward to. We will also be pushing resources to AI and so on, but the surgery team is just one of our development teams, and it's a quite small team.
It's four people, so it's not a really huge one. And in our roadmap, I mean, we have stuff to develop for even with a higher pace for many years to go, where we can really help our customers become more efficient.
Next one. You mentioned that the cost for the customer is 1%-2% of revenue of our customers for Webdoc. Could you speculate where you see this number moving in four, five, six years if you get these agents that you foresee?
I think that, we have an aim to increase that, up to, internal aims, I should say, over time, to increase it to 2%. By, helping our customers, more and more, creating more time for them, having them help more patients, and then a small part of the savings they make should fall down to us. And we can really have a substantial impact for our customers in helping them help more patients. So, AI agents is a part of it. So far, very little used, but it, there is really good potential for it, especially when it comes to medical notes, what is, the certificate, doctors write for someone who is ill and cannot work.
No, I don't remember the English name at the moment, and so on. Decision support, I don't think is very important for us, as I mentioned at previous calls, because most of our users have very simple medical decisions to make in the private healthcare, and it's also the part that they like, and it doesn't really eat any time for them. So we will focus on administrative parts, and that's where we really can help them. It can save them a lot of time and the things they, they really don't like to do. So it's very valuable for them to get rid of that and to instead add more patients. I think that's the most part, and then I think it's also important that they will deliver an experience that's similar or are AI agentic, so to speak.
That's something we're working on, and we have good proof of concepts on it. We'll see exactly in what pace we develop and release what. We'll focus, as always, on what saves the most time and creates the most value for our customers.
Great. Final question, could you comment on the high consulting revenues you achieved in Germany? Where are they coming from? So, this is related to a upgrade on the connections for the public system. And it is delivered by a partner, and then we charge our revenues, and then we have a cost to the partner, basically. And it is something that has occurred historically in Data-AL, basically every other year, approximately. But it's not really a strategic revenue for us. It has a low gross margin of around 25%. But it's something that we had planned for in our targets, so it was important to reach those. But it's not something that will reoccur every year or every quarter, basically. So that was the final question.
Okay. Thank you all for your questions this morning and for your attention. Just reach out if you have any more questions. We're available on email or phone. So, have a good rest of the day, and thank you for today.
Thank you.