Welcome to Carasent Q1 report for 2026. For the first part of the conference call, the participants will be in listen-only mode. During the questions- and- answers 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 first quarter of 2026. We will start with me giving a company update, and then I will hand over to Svein-Martin for more of a financial update. First, a quick overview of who we are. Our business is to help our customers, who are private hospitals and clinics, to help as many patients as possible. We do this through our EHR systems and related products. We're growing rapidly, this quarter with 16% ARR growth and with highly recurring revenues. In this quarter, it's actually above 95%, and we have a net revenue retention of 111%. Looking a little bit more of the highlights of the first quarter. As I mentioned, we had good ARR growth of 16%, and we continued to scale our business in a good way.
In this quarter, we had almost 80% of the growth translating into profitability. This is the way we've been working the last couple of years, and this is how we aim to continue working. In the quarter, we have quite low consulting revenues, and consulting revenues, they varies between quarters, very much depending on exactly what happens, which type of projects we have in each quarter. What I think is also, and we also have mentioned many times before, consulting revenues are not our focus. That's not what we want to build our business on, and we're trying and working to move consulting revenues into ARR instead. It also makes it easier to convince the customer of changing EHR system by not having a lot of upfront costs. This is something we're working towards.
Sometimes, the projects are of types where you can have consulting revenues, and sometimes the customer is of such type that they expect cost there and can accept cost there, and it's difficult to move them into ARR. We have a major milestone as we speak. We're moving our first paying customers into our German system, Webcur, which was formerly named Webdoc. Webcur was a name that was preferred in the German market. It's a major milestone. We're very much looking forward to it and to start scaling the sales of our new system in Germany. Also, we have some promising traction from Medsum. We have signed up more than 100 new users this month or last month, and I will come back to this.
As I mentioned, signed but not ARR is now at SEK 4 million, and what I think is worth mentioning around that is that all those SEK 4 million are new, so all those are sales from this quarter. The SEK 5 million we had last quarter has already been implemented. In total, with the 16% organic ARR growth and the lower consulting revenues, we have 7% reported organic growth in the quarter and an EBITDA margin of 10%, up from 6% a year ago. Talk about what I think most people are talking about at the moment, AI. This quarter, I plan to talk a little bit about how we act, what we do with AI. There are three areas which I would like to mention. The first one is that we accelerate development.
We believe strongly that as it gets faster and faster to develop software, customer demands will increase. What our plan for that is to use all efficiency gains into producing more functionality for our customers and a better user experience for our customers, instead of saving money. The aim is to really increase the pace, get more and more functionality out to our customers. What that means is that we're widening the gap between our product and the legacy systems who do not really develop. If you haven't really developed, AI will not help you very much starting to develop. You don't have the culture, you don't have the structures, you don't have the staff.
I think that will really help us, but I think it also will be an expectation of customers, and it will also mean that we keep a gap to potential new threats and competitors, keep a big gap between us and them by continue to always develop and increasing the speed as we go. We see quite big gains from AI in development and in many other functions also too. At the same time, it's very important that we continue to live up to all regulations, that we make sure the product is always stable and that it's secure. It's a bit of a balance act, but I think we're taking big steps forward and getting better every week and increasing the pace.
We also do a lot with AI in the product, and there's a lot of potential to do a lot with AI in the product. That comes from healthcare has a lot of administrative demands, and it takes a lot of time from our users. AI is really good at administrative parts. Medsum, which is our ambient scribe, that means that we listen in on the discussion between the doctor and the patient and automatically proposes a journal note already pre-filled into the system, where the doctor can then make some small changes if he or she feels the need to do that. What's been very important there is to match the quality of the leading ambient scribes that started a bit before us. We now feel that we are at that point, and I know that many of our customers feel so.
Most of the customers we gain, we gain from other ambient scribes because they already know the structures. They know they like to work with those type of tools, and they just feel it's much better that it's completely included within the EHR system instead of a separate system. The potential gains from this are really big. The studies out there say that you save between 15% and 33% of the documentation time, which it's a lot of time for our users. I will get back to next steps, what we'll do on the next slide, what we plan to really do.
What I think is what we're focusing on now is to really rebuild the UI of mostly Webdoc, our largest system, to make sure that the AI experience fits the UI experience, and it creates, in total, a really good experience for the user and a really efficient experience. We are, this year, rebuilding quite big parts of the front end of Webdoc to make sure that AI is a natural part of it. What we want to do with the AI tools are this, parts of it we already do. When you have the discussion with your doctor, we listen in on the entire discussion. From the discussion today, we just propose clinical notes. The next step is to really catch everything that you agree with the patient to do, prescriptions, referrals, follow-up bookings, sick leaves, and so on.
All of these are tasks that the doctor, or physiotherapist, or so needs to do either while in talking with the patient and quite often do after talking with the patient. It's a lot of their time, and it is stuff that they don't really like. We catch it all so you don't have to remember it. You can just see it popping up in the system as you speak, the tasks that you have to do. While you're done, you just go through the tasks that the AI has proposed for you, and it's as much as possible.
There are some legal hurdles on some of these tasks, but most of them we can have completely pre-filled proposals, both from the talk but also from previous medical notes and other information we have in the EHR system to really make sure that we give such a good proposal as possible. This we aim to have in place by the end of the year. The reason why it takes to the end of the year is that we're redoing the entire UI. We are interviewing, we're sitting with a lot of users, testing different ways of building this, testing how they really want it to work. Because as I mentioned before, the most important is that we build technology that really fits how our users want to work and fits their every day. We are in no stress.
We have the strong position we have, and this will not, in Sweden and in Nordics and Germany, grow exponentially. This will be a steady step that users want to start using this type of tools. We're doing it really properly. We're doing it fully built-in as a natural part of the system rather than just hiring it out. We want to be a fast follower in general. It will be a very interesting autumn, I think, releasing all this functionality. Looking a bit more at the journey ahead, we aim to continue as we do. With strong organic growth, especially focus on ARR, we aim to continue have a good grasp of the costs, not to reduce costs. We believe that we need to release even more functionality to our users. There is so much we can do for them.
We have a tight cost control, and we aim not to increase our cost and that the absolute majority of all growth should translate into profit. This year, we're also launching Webcur, formerly Webdoc X, in Germany, which is a major milestone for us. With that first, I will hand over to Svein-Martin.
Thank you, Daniel. Starting off by looking at the financial highlights of the quarter. The headline growth figure, as Daniel mentioned, looks a bit lower than we're used to, and I will get back to that in more detail on the next slide. The underlying metrics that we follow that we believe drive value in the long term, the ARR growth, the scalability, and the cash flow generation continued to develop in the right direction and improved a lot compared to last year. We ended the quarter with SEK 360 million ARR, contracted ARR. It grew 16% year-over-year, and we had a strong net retention of 111%. Profitability also improved a lot, and EBITDA margin was 20%, and EBITDAC margin was 10%. Starting with the P&L, you see that the revenues came in at SEK 90.4 million compared to SEK 85.2 million last year.
This, as you see illustrated here, is a growth of 6%, 7% adjusted for currency. You see that there was a drop in the consulting revenues from SEK 8 million to SEK 4 million, so a 50% decline, and that is the reason why the growth is lower than the ARR growth. This is a few different reasons, but the consulting revenues typically vary more from quarter- to- quarter and is more lumpy in nature. We also had, in Metodika, we had to focus a lot on post-implementation work for Volvat that didn't generate the revenues. In general, we prioritize away consulting in new sales processes compared to the recurring revenues. Looking at the cost base, which I think is one of the key highlights of this quarter, the cost base is essentially flat year-over-year. You see personnel expense increased SEK 1 million. Other OpEx decreased around SEK 1 million.
The capitalized development increased SEK 500,000 . Around SEK 500,000 increase in total. This is even though we add quite a lot of cost for AI-related features such as Medsum where we have a big hosting cost before we get a lot of revenues, and also in our own development, we add tools. We have been able to absorb these costs and become more efficient in other areas, basically. This results in EBITDA of SEK 8.6 million compared to SEK 4.8 million last year. You see that the revenues grew approximately SEK 5 million and EBITDA grew close to SEK 4 million. We were able to convert close to 80% of the revenues into profits, even though we have very low consulting, which typically has 100% drop-through rate. This is a very important metric for us.
Also worth noting that there is no adjustment in the figure, and there was no adjustment last year either. Looking at the ARR bridge, it clearly shows what the growth is coming from. You see on the left side, we had SEK 320 million contracted ARR last year. The net upsell this year was very strong, SEK 45 million or 15%. This is also stronger than historical figures, and this is because some of the large new units we have added is included here, such as Medtanken and Volvat, adding SEK 10 million alone. This is included in net upsell because we had some units from these customers before we signed the big contracts. It's an existing customer. Upsell in general also was very strong in the quarter. You see churn here, SEK 10 million, 3%. A third of this is coming from the German legacy product.
In the Nordics, we can see now that the churn has normalized again, as we had communicated in H2 that we saw that churn was down to lower levels. If you remember, March last year, a lot of the churn we had in Webdoc there took effect in March, so now this is not included in the figures anymore. On the new customer side, we had SEK 50 million in ARR added, a 5% growth, and it improved quite a lot compared to last year as well, when it was SEK 11 million. In total, ARR grew 16%. Worth noting is that the Nordics is now growing at very strong levels. Excluding Germany, the ARR growth was 19% in the quarter, which is higher than we have seen actually for the last few years.
This is a very good sign that the momentum in the business is continuing. This slide we have shown many times now, and it's a good illustration of how we think about developing the business. To explain it to those who haven't seen it, the orange line there is the revenues, and then the bars is the cost base that we have. The dark blue part of the bar is the OpEx, CapEx, and personnel expense for the organic business if we exclude the acquisition we made. You see that it's been extremely flat over the last 10 quarters. This means that we have also been able to take out costs given that we have wages increase and inflation, et cetera. This is, of course, very powerful.
If you look at how the margin has developed because of this, we had the EBITDA of -30% three years ago. It's now +10%, and we still have capacity to continue to scale, and that's the plan going forward, to add very little cost. We will add some roles this year in Germany to roll out Webcur commercially, but in the Nordics, we will continue to be very disciplined. Finally, on the cash flow, you see that in the quarter we had a very strong free cash flow, 47% operating cash flow and 36% margin on the free cash flow. This is, of course, because of a very strong working capital effect during the quarter, and it's driven by three main effects. Firstly, as I talked about in Q4, we had a big customer that had a delayed payment in Q4.
This came in in January, and it was SEK 10 million added. This would normally have been in Q4. We had very high consulting revenues in Q4. This was paid in Q1 as well. This, of course, boosted the working capital compared to normalized levels. In Q1 is typically also quite strong because we have annual invoicing of certain customers. It's worth noting also that our working capital profile is very supportive to our cash flow in the long term, because we charge upfront and we generally pay our invoices after the period, and salaries. All in all, final point there, share buyback, we did SEK 40 million in the quarter, and it took around SEK 40 million in cash.
All in all, a quarter that in the underlying metrics, the ARR growth, the conversion of revenues into profits, was a step in the right direction and we continued to deliver on the plan on these important metrics. With that, I 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 Elvin Rolder from DNB Carnegie. Please go ahead.
Yes. Thank you. Good morning, Daniel and Svein-Martin. I hope you can hear me and I hope you are well. I have a couple of questions from my end covering a bit of a range of topics. Maybe we can begin on the consulting part, given that it was kind of a contributor to the lower revenue here in Q1. You mentioned in the report that especially the Volvat effects are expected to subside here in Q2 already. Can you kind of help us in what kind of base shall we use going into Q2 here now? Do you think Q2 2025 is still very unrepresentative? You mentioned the comp in Q1 was difficult as well. Can you help us kind of bridge that a bit to understand that?
Yeah, sure. I would say that the Q1 is a bit on the weaker side on consulting than like a normal level because we didn't have Metodika doing a lot of consulting work. In general, it's not the key priority for us as we mentioned. It's not that simple to be very specific on guiding for the next quarter and the year. I would say in Q1, it's a bit lower than what we have as a run rate level when Metodika is fully free to do consulting work. I don't know if you have anything.
All I could say is that what happens is that Metodika, with the big implementation of Volvat, we had some work to continue to do in the first quarter, which we do not charge them for at all. We feel it's very important to have a happy customer. We can also see it, we get a lot of questions from potential customers in the Norwegian market from successful Volvat project. I think that's worth the investment. They will move on now to more of a normal business in the second quarter. We should really see consulting revenue picking up. I can also mention that even if our focus is ARR, we want to, of course, have as good result as possible. We're also looking over a bit how we work.
It used to be that when we sold to a new customer, we had not high fees, but we charged them for education, setting it up for them and so on. Now we very often, to a very large degree, give it away to the customer as part of the negotiation and not to negotiate the lower ARR or the subscription base. We don't want to lower that. Maybe we're giving it away a bit too easy. That's something we're discussing internally. I don't have an answer for that yet.
That's something that we're looking at, because ARR is our priority. Of course, if it's possible to also, without hurting the subscription revenue, to have some other revenues, that's fine. That's something we want, of course. We're looking over that a bit, if we're giving it away a bit too easily. In general, the coming quarter will be better than this quarter, but not as good as the best quarters last year when it comes to consulting revenues.
Good. Thank you. Very clear. Continuing maybe on Germany. You attributed 1/3 of the overall churn in the group to the legacy products here in Germany, and we saw a decline in revenue overall for the German operations here in Q1. Should we expect revenues from Germany to decrease for the full year 2026? Do you think that you will be able to compensate the loss from Data-AL in what you gain on Webcur? Is that perhaps too early?
Yeah, I think that's a bit too early. This year it will probably continue more or less in that trend, I would say. The major thing to look at this year is to really get Webcur up and running, happy customers, the first ones who can be references for future customers, that we get sales and marketing really going and so on, for them, during next year and the year after that, to really see growth coming and that we're growing much faster than we're losing customers. It will take a bit of time.
Yeah. Thank you. Very clear. Moving on to Medsum here. I believe a couple of quarters ago you mentioned that you have an ASP of, I think it was SEK 750 / user per month. Please correct me if that's wrong, but is that still a kind of a valid level to assume? Has that had an effect on sales yet? You mentioned 100 users live or signed since last month already. A bit tied to those users, is there anything exceptional about this usage that we should be aware of? Have these been involved in the pilot programs, and have been involved since the start in this? I guess the essence of the question is, should we expect the next 100 customers to come as fast, basically?
That's a good question. It's a bit hard to answer, actually, but I'll try to give as much information as I can. No, those were not pilot customers. We had roughly around 50 customers through their pilot phases, the last maybe six months, a little bit on and off. Roughly 50 customers, they are still there.
Users.
Users, I mean. They're converting to paying customers. Since we then relaunched, we have gained another 100. When every Monday we go through sales, and if we look at upsales to customers, Medsum is now quite a large part of all discussions and all the meetings with customers, and also in signed. I think that's going really well. I think it's a bit hard to predict. The ones we win usually now move from other such solutions. Those are early adopters. They've used these type of solutions before. They like it. The rest who has not started using this type of tools, how quickly will they start using it is kind of the question. In general, this business is slow. I mean, healthcare workers are trained not to take risks, and they are a bit conservative.
Our group of users are a bit more offensive. They're on the private side. Still, they have the legacy of being educated and working in the public system for their entire education and far beyond. In general, they are slow. This technology, on the other hand, has really potential to save a lot of time for them, and it does not force them, at least not to a large extent, to work in new ways, which is typically the barrier. I have hopes that this type of technology will go faster than other technologies, and it will also be more enduring, I mean, that they stay. I think many of us were in the time where video meetings started, and it was the craze, and it kind of didn't really grow very much, and now it's declining. This, I think is different.
I think this really saves time. I think it's easy to start using. For many of our users, it fits really well. You should also keep in mind there are other good solutions for it. We have in our systems, phrase libraries. Many of our users have very similar patients all the time, so they have already pre-filled what text should be there, and then they maybe change from right to left eye or something like this. It's not for everyone, but I believe it to be for many of our users, in the majority over time. As we add more functionality during the year, I think we'll see more and more users that say, "Okay, now it really saves a lot of time for me too." It's a bit hard to predict how quickly will users pick it up.
Just mentioned there on the video is that one challenge we have, but we also overcome in these numbers as you can see, is that the digital care base, they are actually declining a bit in Sweden, and most of them use our systems. It's a bit hard to predict where technology is going. In general, I think this one will pick up quite rapidly compared to other technologies for healthcare. That was a long answer. Sorry for that.
Yeah. No, it's great. Thank you, Daniel. A couple of more from my end. Sorry about that. Perhaps, can you give a bit of an update on the surgery module? What have you seen here? When should we expect this to kind of move the needle in terms of contributing to the ARR, or how much has that impacted ARR already? Can you give just some kind of general update here?
Yeah. We are in a lot of those sale processes, so that's the positive side. All of these customers, they are 100+ users in each place. We don't charge them per user, we charge them per visit, but it's still a good indication on the size, and 100 for us is big. Our average customers, I assume, is eight to 10 users in Webdoc. Those are quite big, and you're like a small hospital. It's also then extremely slow. I was sitting in such a meeting with a potential customer not long ago, and yeah, we will move into Webdoc. We believe that this will happen maybe in two years' time somewhere, and it's very frustrating to sit there and have those discussions. It does take time.
They want to go through, they want to make sure they have all the staff aligned, that they plan time for it. They want to make sure they do the risk evaluation. That's their legal requirement on them to do. They want to go through every single part of the system, make sure that they have all the processes in order, and that it will support all the processes they have. They want to make sure that they then change if there's something that they need to change in the process. We want them to change those processes so that they really are efficient, because quite often they work with systems that don't really support the process today, and they've had to build processes that are not optimal, but the system has forced them to do it. It really takes time.
I know it's not what you want to hear, it's not what I want to hear, but I really believe that it will support strong growth for many years. It's not coming as quickly as I thought and hoped. The discussions are really positive, and we have an increasing number of users in the OR module, I would say also, all the time. To move the needle, it seems it's going to take a bit more time than I expected.
Okay. Very clear. Thank you. Let's see here. Maybe perhaps on ARR, continuing on that, you have a very strong growth in the Nordics here. I think it was 19%, and you have quite a good kind of new customer inflow. Could you expand perhaps a bit on what is driving this? You've mentioned a couple of, over the past year or two now, that you've been trying to kind of been more active within your sales processes, hiring salespeople, and been more kind of outbound in your activities here. Are we kind of seeing the fruits of that labor now, or is there any other kind of factors driving those figures we're seeing?
Sure. Yeah, I think we are seeing the effects of a more effective sales force in the Swedish market. On the sales side, last year was actually quite a strong year. We had the churn levels that took down the growth a bit, so it was netted out a bit. Compared to what we sold in terms of ARR, and not the reported revenue, but the sales was up quite a lot. It's the different effects that we had talked about a lot last year. We restructured the entire sales organization, which is showing good effects also, where the sellers can focus more on chasing new customers, and we have customer success representatives working on the existing. This has had effect.
We have also had very good traction in Norway, where we have taken a lot of market share from competitors, basically through having invested over time more than them in the product, and also in terms of marketing and working proactively towards new customers and new segments. It's a mix of factors, but I would say that's the main reasons.
Yeah. I think we found a quite good balance between kind of understanding the customer and being part of the healthcare system, but still having a sales approach and being a bit pushy, and that's a balance act. I think in general, that's where we're quite good, I would say.
Yeah, like Daniel mentioned as well, on the backlog, it looks like it's quite a bit down, from SEK 12 million to SEK 4 million. On the other hand, that SEK 12 million, it was big customers that we sold that was in the backlog for years, that took time to implement. This backlog this quarter is basically all that we had going into the year is spent, and then we have added SEK 4 additional million, which is actually quite a strong quarter for us in terms of sales.
Okay. Yeah. Thank you. That's also very clear. I will ask one last question, then you can leave and some other people's last questions perhaps. Perhaps on the CEO letter, Daniel, I think you're once again kind of enthusiastic about 2026. You're looking forward to it with kind of an upbeat tone, I would say. How should we kind of bridge that enthusiasm in terms of kind of the ARR development, expectations for Germany, expectations perhaps in [audio distortionYacht] alarm considering what has happened there, perhaps for Medsum? It sounds like with all of these effects that ARR should kind of accelerate from this kind of high level. Is that completely unreasonable, or what are your kind of thoughts and how should we kind of bridge that thinking? It's a bit of an unfair question, I understand that.
I can try to put some color onto that. I would put it this way. I'm very upbeat, actually, about what we're doing. I see that we every day get much better at shipping out functionality to our customers, both for AI but also through just getting better as an organization. I also see that the AI tools within our systems, I mean, for our customers, will add a lot of value to them. I'm very optimistic about those possibilities and also with the rebuilding of the entire UI, so it really fits, and we create a seamless good experience. I believe very strongly in that, and I also believe in Webcur. Many of the things that I believe strongly in, that we're working hard on at the moment, will not translate financial performance this year.
The most of the things I'm mostly upbeat about is maybe more long-term. Because this complete new UI with the built-in AI, that's by the end of the year, and our customers are slow at adopting new things. I believe it will take time. On the other hand, I also see that, as I mentioned earlier, that the customer representative, especially, they're really good at talking about Medsum with the customers. They're signing up new customers. We have a webinar tomorrow. I think there are 175, roughly, people who's going to listen in to that webinar, who has signed up for it, and it's only about Medsum. I think how to bridge that, if you need to bridge that, is that I think this year we'll see the Nordics growing well. Germany, as I mentioned before, will not really grow that much this year.
Germany, during this three-year period that we have targets, will accelerate through that. At an end of that period, we really see Webcur being a big growth factor for the group. It's Nordics that will carry this year. I think that some of the upbeat and the potential I see is a bit more long-term than 2026. I think that's maybe if you see a difference there, if you see a gap that needs to be bridged. Did that kind of answer it, Elvin?
Yeah, of course. That's very clear, and thank you so much for answering all of my questions. I will get back in the line here. Thank you so much.
Thanks.
As a reminder, if you wish to ask a question, please dial pound key five on your telephone keypad. The next question comes from Fredrik Nilsson from Redeye. Please go ahead.
Thank you very much. Good morning, Daniel and Svein-Martin. I want to start with OpEx. While OpEx, as you mentioned, is rather stable year-over-year, it increased by over SEK 5 million relative to last quarter. You mentioned AI hosting costs, but could you elaborate a bit on the increase further?
Sure. Q1 is a bit higher than Q4 because we have holiday effects in December. That's a part of it. We also said in Q4, right, that we have this SkatteFUNN grant that had the SEK 2 million effect or positive effect on the cash cost in Q4, that we talked about in the quarterly call. That was part of it. It was AI cost that we added. It was around SEK 1 million, I would say. That's the total. As you also saw that it's quite flat year-over-year.
Great, thanks. Regarding Medsum and its typical users, you mentioned, I think it was a few quarters ago, perhaps, that a lot of customers are testing different solutions. Do you see that they are more settling down now for Medsum, hopefully, or what do you see currently among the customers using Medsum or competitive solutions?
Yeah. I believe most of them will settle down for our solution as it's completely built into the system and has access to more information than the external ones. I think it's just a better user experience, and that's why you will settle down for them. The key being at least on the same quality as the competitors when it comes to what we actually propose. That's the challenge. We need to make sure to always improve there. I believe that we are there, and as long as we are there, they should choose the built-in solution for most cases. There are use cases which we do not aim to meet ourself, where we happily work with partners. For example, there are some users who want to walk around and kind of record as they go.
Maybe they're a physiotherapist working a lot in the gym with the patient. We do not at the moment at least aim to build an app for Medsum and try to be really good in that experience too. There we're happy that you choose something else then. We'll see. It's still early days and since most of our users joined last month, it's difficult to answer if they will settle down. I believe that we have a good edge because of our position and what the information we have and since it's the system that's already being used. It's up to us to make sure that they do stay there, I would say.
Okay, I see. Just to follow up on that, you mentioned that a lot of your Medsum customers are coming from other solutions. I assume then you don't see any significant movement in the other direction, Medsum customers moving to your competitive solutions, to the competitors, I should say.
No, not yet. We did see it early on when we did lack the quality before we started selling, because when we test ourself, if you just do simple meetings and so on. Also in those test cases we had with early pilots, it looked to us that we could match the quality even when we used smaller LLMs. We thought so, but then when it entered the real world and people started talking about golf through the meetings and there was a bit scrappy sound and things happened, our quality was a bit behind, and that's enough for them to choose another solution because then you have to do some extra work with the keyboard and they really don't want to do that.
You should also keep in mind that what they do in general, or what they want to do is kind of not to take notes during the meeting anymore at all. You have to make sure that you catch everything that they want to have in their notes, because otherwise they anyway have to take notes themselves, so to remember what they want to write afterwards. It's really sensitive on the quality. We paused the sales and then we had some customers leaving and choosing other solutions. Since the restart of the sales, when we're using the larger LLM, we have not seen that. It's early days.
I see. Thanks. Lastly, some years ago you used to say that your customers have a very long list of new feature they want, but you have to focus on the most value-creating ones. Has that changed given the advancements in AI? Could you make those easier now? I suppose you can, but at the same time, there's also a lot of new features related to AI and the UX that you want to invest in. Is that something that you will focus on more, or is there other things that's more interesting?
Yeah. We still have a very long list of what we want to do for our customers. Extremely a lot we want to do and that could really add value for them, and we keep having those projects competing with each other. You have, as you've mentioned, the new AI functionality that also in there competing. At the same time, we're increasing pace, but what we say internally, which I think is a good point to do also here, is that even if we get 500% more efficient as a total, in some things we will be much more than that, but in total. That's not a prediction, but that's just an as and if. We still have things to do because the systems we are supporting, they are huge, our systems.
We're supporting a small hospital and a clinic in some cases and quite large hospitals with everything they do, all parts of it. It's our systems compared to other types of products that you use are not as polished. They are not as good in all respects. I think that AI will start demanding that of products like ours too. There's really a lot we can do and need to do and which will add a lot of value for our customers. I'm really looking forward to that. I believe that even if we get 5% more efficient, we will have jobs for those that work for us, as many jobs.
It's only conditional on that you are prepared to change, that you're prepared to really embrace the new technology, embrace that your role will change quite a lot, and then you have a place. That's what we say internal, and I think that's the truth, is that expectations will go much higher. If you take the old list of everything we felt that we want to do for our customer that will add value, we have a lot of new demands too. I think that's where we're heading, but it's also difficult to know. Things are moving quite rapidly. They might not come to the level I expect them to, or they might go beyond it. Who knows? We're trying to really make sure that we are the winner in all those scenarios. That's our way.
That's interesting. Thank you very much.
Thank you, Fredrik.
Thank you.
There are no more questions at this time, so I hand the conference back to the speakers for any written questions and closing comments.
Okay. We have gotten quite a few questions here in the chat as well. First one, how much has the hiring of new salespeople affected the revenue?
I think it's a bit early days. Most processes are quite long. When we go through the list of potential customers that they are working on, I feel that we have good progress, good discussions. They're not many that we are invoicing yet. Some of those signed, that we leave the quarter with, comes from new salespeople.
Next one. With profitability and positive operating cash flow now established over several quarters, how are you approaching capital allocation priorities? In particular, how do you assess the attractiveness of reinstating a share buyback program at current levels?
That's really a Board decision and AGM decision. We believe in buybacks in general as a way of, instead of dividends, as returning cash to our shareholders. That's what we want to do. As also mentioned, we are looking at a few acquisitions. Not in general, we're never in any structured processes, but we have a few acquisitions that we think are interesting that really strengthens our position in Sweden, key acquisitions. They are always difficult to know when they will happen, as it's us knocking on the door saying that we're interested, not the other way around. In general, we see share buybacks as a good way, and we see that we'll have an increasing cash flow from operations, and it will continue to improve. We will use that for share buybacks and those targeted acquisitions.
Yeah, you see now that ending the quarter, we had a cash balance of SEK 130 million. We don't plan to have that large of a cash balance in the long term. We will deploy it either through buybacks or these strategic acquisitions that we have mentioned. Next one. What are the opportunities in the pipeline to continue growing over the midterm? Signed not implemented ARR currently stands at SEK 4 million, which is low compared to previous periods. Yeah, I think we addressed the backlog there a bit earlier. Over the last few years, we had a big backlog, but it was also customers that was in the backlog for quite some time because it took time to implement, and this is basically a new backlog, which is a very strong quarter for us, I would say, signing SEK 4 million new ARR.
In the midterm, we see great potential to continue to grow. We still have a low market share in the Nordics, much market left to grow in and to take market share, and also in Germany as we get more traction there. You see that momentum in the ARR this quarter already with the 19% growth in Nordics, which is very strong, I would say, for us. Next one. How will you attract new customers in Germany? Through salespeople on the ground or company partners?
Yeah. That's why the acquisition was important to have something to build on. We have a sales force in Germany. It's small, it will have to grow. Working, and we're also selling through partners already, the existing products in Germany. I think we'll continue with both. What we really want to develop in our German organization is much more creating leads through marketing. We still work very conservatively into Germany with a lot of cold visits, and that's quite an expensive and inefficient way of selling. We have to be mindful that Germany is not like Sweden. We try to do cold visits in Sweden, that really doesn't work. No one wants to meet you at a cold visit. It is a bit different within the markets, and we have to appreciate that.
We are strengthening the German organization with local experts on how to sell and market in a modern way in Germany. That's something we are in the process of strengthening the organization. I think we have a good base to build on with good salespeople that we can then help become even more efficient.
Okay, great. Next one, could you please update us on your progress in Stockholm, also with regards to further hiring of further sales reps?
Stockholm is moving to Cambio from Take Care in 2029-2030. I think you can all read in the papers how those regions that moved to Cambio feel. There are a few that actually moved from Take Care. As a user, you feel that Take Care is a much more modern system than Cambio. It's also much more suitable for small private clinics than Cambio is. When that change happen, it will really help us grow a lot, is our expectation. It is still so far in the future that it doesn't help us very much in Stockholm because we can say you're going to change systems anyway, but typically this process get delayed also, so it's five, six years in the future. That doesn't really count that you anyway have to do it in five, six years' time.
It doesn't really help us from that perspective yet. Still, Stockholm is the biggest market we can win customers from Take Care. We can continue work, but over time, that change we believe will really strongly help us. It's not yet. Still, we're growing in Stockholm. Most new customers are in Stockholm. We will not hire new sales reps at this point in time. I feel that the ones we have hired are still catching up and coming up to speed, and it's the same with the customer representatives, customer success. They are still kind of learning their roles, but doing it at a really good pace.
Next one. Could you comment on how the UI should look like in your perspective in an AI world? How will that incorporate workflows in conjunction with AI?
Yeah, that's something we're really testing out and trying out. We have a lot of folks that we are testing with real customers, not with real data yet, because there are limits to what we can do then. We do a lot of videos with them testing the system and some while we're interviewing them. I don't think it will look, in the first version this year, extremely different than from before. We're rebuilding the UI, and at the same time, we are addressing all things we want. We know we anyway want to improve. Everywhere we're building in AI support so that if we have the knowledge from the meeting, from the recording, or from anywhere else in the system, we pre-propose things. There are limits to what we can pre-propose.
We are currently not allowed to propose things that help them make a decision. It's more of the administrative side. I think that's quite okay because that's where our users have a lot of work to do. It's not the actual decisions they make. They're typically quite easy in the private healthcare setting, and not what take the time, and also the administrative part, they really don't like. There are a lot of legal hurdles there when it comes to prescriptions. There are a lot of requirements, and we have to follow the exact same look as Läkemedelsverket has already approved in our system. There are certain limitations within the sector, but in general, we think it will not look that different, not at least yet.
You will feel, and you will see the AI supporting you in all different views in the system, and it will also take you to the right views and the right places when you need it to. We don't have it all in place yet. We have attempted how we think it will look by the end of this year. It's still being worked on with the details, but then it will continue to evolve, I'm sure.
Next one. Could you help us understand how difficult it is to switch a system currently, and which part makes it hard or easy?
I think that the hardest part is that healthcare, if you run a clinic, you have all these fixed costs, and I know I talk about this often, but you have all these fixed costs, and then you have your variable income. The entire mentality among our customers is that every day we need to fill it. We have to make sure that we don't have a single empty slot because then we're not making any profits. That mentality means that you're really working the same way. You don't really take a step back and think about, how should we work with our process? How can we improve?
I think that's the major shortfall of European healthcare, in most European markets at least, where you don't really think through your process and think, okay, if we do this work now, then will every day get a bit more efficient? The same thing with changing systems. They see the short-term pain that comes from that you need to re-educate your staff. You need to do these assessments of risk, of changing systems that you have to do as a legal obligation they have, and we support them with it. You also have to work through, okay, how do we work now? In general, we try to support all work processes they have today, but quite often they come from them adapting to a not great system, and they should really rethink a bit how they work.
It's difficult for them to take the time to do it because they have such a short-term focus when it comes to planning and everything. I think that's quite a big hurdle. We're also a bit hurt, makes it a bit more extra difficult, I would say, is that when the regions change systems and you read everything about it in the newspaper, it makes everyone more afraid of changing systems. Even though when the regions do it, they do it in a completely different way than us, so it is not that difficult. I think many have the view and the feeling that it is very difficult, when it is not that difficult when we help them.
We ask all customers a month after they change to Webdoc, "How happy are you with the change?" It's a scale from one to five, and every month we are above 4.5. Already a month after, they're very happy with the change. We try to mark that, we try to tell that story. It's a sector that's afraid of change.
I think we got the follow-up here on the AI discussion, where it says, "Does this mean agents could be deployed? Are you planning to connect MCPs? How does this work with regards to restrictions on sensitive patient data?"
Yeah. Those are tricky questions. In general, we are not allowed to train on sensitive data, but we can give them access to sensitive data through MCP, for example. That's fine, a bit depending on what you do. It cannot be decision supported within the new certifications, which we might do. We haven't decided yet. One reason why our AI cost gets quite high is that we also have to make sure that it's always processed within Europe, and that's from some of the large LLM models that doesn't cost extra. When we do quality checks and we do blind tests with different LLMs, we can change in a day. OpenAI is the one that scores highest in blind tests, from Google, Anthropic, and OpenAI. Though the OpenAI charge you quite a lot for guaranteeing that it's within Europe, and also guaranteeing that it's zero retention.
Zero retention is not a legal requirement, but it's something that's really appreciated with our customers, that the data is never stored in an American-owned cloud and if it's within Europe.
It's worth noting that they charge a lot. It's like a minimum fee.
Yeah.
It's not like the gross margin will be very weak. It's more that we need the big volumes.
Yeah. Because the token is better, and it's lowest with OpenAI of all of them actually. OpenAI uses less tokens than the other models for the same query when it comes to these questions. Over time, that will make us more cost-efficient. In general development, it's mostly Claude we use, but in the customer-facing products, it's OpenAI at the moment. That's easy to change. Let's see, did I answer the full question? Yeah, the sensitivity. Yeah, it's difficult, but as long as we stay within Europe, we do not train on data, then we're fine. We do use the feedback. When users change something in their medical notes we have proposed, our structure learns from those changes, but we do not train the LLM model. We set up the structure so for all changes, it learns from it.
It gets more and more personalized because each user wants their replies and their medical notes in a bit different way. It learns from that. We're not training on the sensitive data. We are training on other data.
Next one, how does Medsum differentiate itself from competitors like Tandem Health?
We're not really trying to differentiate ourself. We're trying to be on the same quality level. Differentiation factor we have is that it's completely built into the system and have access to all the data in the system. It's one user experience. I think that's how the difference we want to make. I think Tandem is good for certain cases that we do not aim to meet because they cost too much for us at the moment, at least to develop compared to the amount of our users that need it. They're going a bit of different direction, I would say, than what we're doing, and I think that's fine and that's good. We have a good collaboration with them. For the typical Webdoc and user of any of our systems, I think that Medsum will resolve their needs in a good way.
Next one, is Medsum a transaction-based revenue or a subscription from a segment reporting standpoint? It's a subscription revenue that we charge per user.
That have been said, we want to move from charging per user over time. That's just because everyone we're competing with are charging per user, but we want to charge for usage in the future, like we do with Webdoc, charging per visit. I think that makes more sense. As they use it more and more, we'll have more revenues, like we do with Webdoc.
Yeah. The final one, I think you addressed it, Daniel, but can you comment on recent sales development in Stockholm and VGR?
Sorry.
Can you comment on recent sales development in Stockholm and VGR?
Yeah. Most of our new sales are in Stockholm. That's where we do most of our progress. I think it's also extremely important from a long-term perspective that we get more and more known in Stockholm. When they're changing systems, we are well-known and everyone thinks about Webdoc. We have to make sure that that's the case. Also here now, it's the biggest market. That's where everyone is and that's where we focus our sales efforts. In VGR, we are continuing sales also, but it's a smaller market and we have a stronger position already, where we do have a lot of discussions are within primary care. We're getting a larger and larger part of the private primary care market in VGR.
Okay. That was the final one.
Excellent. Thank you all for listening, and if you have any questions, just let us know. Thank you.
Thank you.