Trifork Group AG (TRIFOR) Q4 FY2025 Earnings Call Transcript & Summary

Dear audience, welcome to the presentation of Trifork's Fourth Quarter Results for 2025. My name is Frederik Svanholm, Group Investment Director of Trifork. Today, our CEO, Jorn Larsen; CFO, Kristian Wulf-Andersen; and COO, Charmaine Carmichael, will be providing a presentation of approximately 30 minutes, followed by Q&A. We have 1 hour in total. Before we start, a couple of practical information, first I would like to inform you that this presentation is recorded and will be made available in its full length on our investor web page. Second, I would like to inform you that if you want to download the slides for today's call, you'll be able to find them on the front page of our investor website. Third, we invite you to ask questions and engage with management after the presentation. Before we get started, we have to present this disclaimer. Okay. Let's jump to the presentation. I now hand it over to our Group CEO, Jorn Larsen.

Thank you, Frederik. And before we get going, I'd like to start with a few opening remarks. And so first of all, we will be three presenters here today. "Oh, actually four," Fredrik also counts, but there are three faces on this page. And besides Kristian and myself, you might have met before, we have Charmaine here, our new Group COO. And she will also introduce herself a little bit later in the presentation. She will take part of the presentation. And I just want to say a few things about this picture in particular because actually, right now, I'm sitting on -- in this building, our Trifork building. And it was not me choosing this picture, but I find it interesting because you also see a plane. And I don't know what you see, but I see this is a Swiss plane, and that means it's flying with our software. And maybe I was even in that plane landing in Copenhagen for this presentation. So one fun remark I could mention, and that made Frederik nervous when I said I had a little surprise remark that is not in the presentation because we always talk about growth and percentages and quarters. So I just want to mention one number that may sit with you well, and that is we actually have 30 years of anniversary this year. And the first year, I remember clearly our revenue. And our revenue in the first year in '96 was EUR 30,000. And I haven't adjusted this for inflation or anything, but that's actually a 7,500x up to our reported '25 total revenue, which is a record and considerably higher than the first year. So I just thought that would be a little bit reflective to mention that. So let's move on, Kristian. We have a lot of good things to talk about today. So today is a good day. It is with relief that we are presenting today. 2025 was an intense year in the world. There are many intense things going on in the world and everything that is going on in the world has some effect on our business. So I remember in summer when we were presenting our Q2, I remember the analyst and investor remarks to our numbers and the guidance and everything. And it's with great relief we can say that we made it through very well. And as we will present in this presentation that we are on a very good track. So if we look in Q4, in particular, Q4 was a very good quarter for us. So the product-led strategy that we set out a few quarters ago is really working. We see now in Q4, in particular, that we had a 27% growth in product revenue. We have high demand for sovereign data solutions. We have high demand for our AI products. We have high demand for our healthcare platforms, and we see good traction in aviation before and hence, the picture from before. We see that in public sector traction that things are going well. We have won a lot of framework and contract within public sector in Denmark and other places. And particularly, I want to mention that we have -- we are building a new product for psychiatric care. It is so that the mental state of the world population is not really improving and a lot of young people feel stress, lost or mentally unstable. And it's very important that we do something for the young generation. And this is a product that will help the digital system, the digital healthcare system to help these young people get back on track before things get too bad. So I'm very proud that we are involved in this. Thirdly, I want to mention that we have a lot of cost control activities in '25 and in Q4. And we see that also has a good effect on margin improvement. We actually see a 3.5 percentage point margin improvement in Q4. We have a strong cash conversion, but Kristian will talk about that later. We also see that Lab is continued to support the core business of Trifork. We have exited some shares in one of our Lab companies. Based on that, we are announcing a EUR 10 million buyback program. We already completed a EUR 2 million buyback program here at the end of February. And it's the way we give back to our investors. It's the way we say thank you to our investors. Also, fourthly, I want to mention that we are also becoming a company with a stronger employer brand. And you can see here one of the latest surveys where we are actually ranking second most attractive within Consulting and Services in Denmark in between engineers. But let's move on. Okay. So 2025 was a record year with as -- of a total revenue of EUR 221 million. And for the year, we had a 7% organic growth and for Q4, a 9% organic growth, and that's a good trend that we'd like to see continue. And also, we had margin improvements. But what we also see here is that if we look towards 2026, which we have also indicated here, is that we see a -- in average, I would like to see a 10% revenue growth for 2026, but we are aiming actually to see an even higher improvement in margin, actually a plus -- yes, between 16% and 32% growth in margin we like to see. So let's move on. Okay. So here, we see the world map. We see that we are active in Europe. We are active in U.S. in business. And then we have a good customer or we have a good market in the Middle East called Oman. And we are helping the Omanian government to employ local people in our Trifork Oman to digitalize the Omanian society. So I'd like to say that between Denmark, Switzerland and Oman, we are helping each other to accelerate digital implementation and the good use of AI. And I also want to just highlight one logo here in the middle of the picture, a little bit to the right, Cellnex. So Cellnex is one of the biggest companies in Europe, if not the biggest with plus 150,000 real estate for putting antennas for our mobile network. Everyone is using the mobile network every day. And without a company like Cellnex, this would simply not work. And we are very proud to be a vendor to Cellnex and a partner to Cellnex. So let's move on. Good. So what has driven the good numbers and the good traction? And I would just mention four things. There are other elements, but there are four very important ones. First of all, we see very good traction in sovereign data demand, and we announced very recently a new data center, and there will be more to come. And we like to help local governments, local companies, whether it's in Denmark, in Switzerland, in Spain, in Oman to have their data under their own roof or very close to where they are. It can be in our data centers. And we use our own software there. So Contain is a software layer that makes us able to host and manage data in a secure way that could actually also be run in a hyperscale. So we can very dynamically move things on and off. XR adoption is still strong. It was one of our, you can say, leap innovations when it came out from Apple, this headset device. But now we see serious applications being built that help enterprises reduce cost and risk, and so we are working, among others, in airlines and the train and transportation industry to reduce the risk of accidents in transportation. Then I want to mention also the adoption of Enterprise AI. There is a lot of hype around AI in these days and months and years. But it's not something that just happened overnight. This has been going on for years and years and years. But now everyone can start up and use some kind of AI and they can clone themselves, and that's all wonderful. It's also a little bit scary. But what we do is that we help our customers to take use in a safe way of their data because we have to remember that AI is trained on public data, but the most data in the world is actually behind walls. And so we enhance the AI models with the enterprise data in a safe way. And then also Public sector efficiency. We see a big potential in growing in public sector across our markets and especially led by digital health, but also other institutions. And I will come back to a story a little bit later. Let's move on. So what do we actually do? It's important that we keep in touch with reality, and we explain to you what we actually do. So for a number of years, we have been working with public institutions in Switzerland, the cantons. The cantons are responsible for education and schools. And the public school system in Switzerland is very strong. It's very well funded. There is also a lot of private schools. You might think that living in Switzerland, everyone flies a helicopter to the private school, but it's not the case. So our employees and colleagues in Switzerland, they have their kids and they go to the public schools. And examination in public schools are done to a vast extent with our software, and it has been going on like this for a long time. And as Andreas Walter expressed himself, he says that we are helping them to transform data to insights so we can improve the education for kids going forward. And this year or last year, we introduced an AI element to this because when you write an essay, it can be fairly difficult to score it in an objective and fair way. And now we are using AI to be more objective when we are scoring these assays. Let's move on to the next case. So here is an example of how we have helped 43 Danish public organizations to support a system that is doing case management and it's used by 50,000 daily users and it's 100% run and operated in our data centers. And that's one thing we would like to help a lot of other organizations with. Let's move on. So finally, here of the case stories, I mentioned already Oman on the map, and I tie it down to here that we are helping the National Institution of Digital Health and we're creating more fairness, more transparency and faster care in Oman. So this is a system like we're doing in Denmark for the whole population, 5 million people; 2,000 institutions will be connected through the system we are building and implementing at the moment, and we are already several milestones down this road. So let's move on. So here, we see -- actually, I will now hand over the word to you, Charmaine. I almost got carried away.

Good morning, everyone. And just by way of background, Charmaine Carmichael. As Jorn said, I'm the Group Chief Operating Officer. Many of you I haven't met yet. So just as a brief intro, I've got an international background in leading and scaling technology organizations. My focus is on operational performance, commercial discipline and as we will talk about consistently through this presentation, product-led growth. My experience includes the organizational development to support that, how do we strengthen go-to-market execution and how do we support sustainable, scalable growth across what is increasingly a very complex and multidivisional environment. So by way of background, let's jump into the detail on this slide. In terms of revenue mix, I'll begin with the revenue composition. In 2025, our product revenue increased from 27% to 35% of total revenue. So that means a larger share of our revenue now comes from our own products rather than pure consulting engagements. Services do remain an important part of our business at 65%. The change is not about reducing services. This is about increasingly connecting our services to our own platforms and products because this supports stronger scalability over time, and it also provides increasing pricing power. This development across Trifork is gradual but deliberate. We're not accelerating at the expense of either stability or our customer relationships. Our approach is highly disciplined. Our long-term ambition is about reaching around 50% product revenue, and that ambition remains unchanged. And we are progressing, as you can see, step by step. Our sector exposure remains balanced, but as Jorn already referenced, we now have 42% of that being public and 58% being private, driven in the main by the adoption that we talked about with tailwind offerings, the sovereign data hosting, our managed service solutions and our public sector efficiency products. So the key message on this slide is a very steady improvement in our revenue mix, while we continue to maintain a stable foundation across the business. Next slide. In terms of customer diversification, if we turn to the revenue quality. Our revenue has grown from EUR 159 million in 2021 to EUR 221 million in 2025. At the same time, we have reduced dependency on any one single customer. Our largest customer now represents 5% of revenue compared to 10% just 4 years ago. And revenue from the broad customer base has now increased to 60%. That shows that our growth has been broad-based rather than concentrated. Our retention remains strong, and that is important within the context of Trifork. We continue to work with 17 of the top 20 customers, and this highlights our customer loyalty. As our product share increases, we expect our lifetime value and retention to further strengthen. So while we are evolving our revenue mix, the stability and diversification of the business has improved. And this combination growth with a lower concentration risk strengthens the overall quality of our revenue. Next slide, please. In terms of direction, looking ahead, our direction remains consistent. In 2026 and 2027, my focus and the focus overall is on execution. We will continue to refine how we bring our products to market. We will continue to strengthen our value-based pricing. We will prioritize our product-focused acquisitions, and we will continue to build leadership capabilities across our divisions. If I could take a second to focus on product-focused M&A and if I pick one of our segments, for example, airlines, what that would mean in reality is that we would acquire either niche software or technology companies potentially that strengthen our existing aviation platforms. And these would allow us to expand scalable recurring product revenue rather than adding stand-alone consulting capacity. As you would imagine, this is not a rapid shift. It's a very controlled and phased development across the organization. As our product revenue increases over time, we expect this to gradually support improving scalability and importantly, margin quality. We are managing this transition very carefully, ensuring that the progress we make does not come at the expense of stability. So overall, Q4 confirms our steady development in mix diversification and operational discipline across the organization. I'll now hand back to Kristian. Thank you, Charmaine, and welcome also in this forum.

I will move forward here with some highlights in relation to the group financial performance. First of all, here with Q4, as Jorn mentioned, we saw a 9% organic growth. And if you look back to the 5% total growth, then the difference here is that the total growth was then impacted by deconsolidation as we published about in Q3 of our Trifork Security business, which now is a Traffic Labs investment. So that's the difference in relation to total and organic growth in Q4. What we saw in Q4 was also a normalizing of profit margins. So overall, for the adjusted EBITDA for the group, we saw an 18.5% margin -- 18.1% margin, which, as Jorn said, was a 3.5% improvement. Overall, if we look into the cash generation, then the operational cash flow conversion was 98% for the full year, latest 12 months, which is a strong position for us, a good number, led by even more disciplined cash collection processes and also by the way that we engage with our customers. Overall, the cash improvements over the year, we saw an operational cash flow of EUR 32 million, also helping us to get to a net debt-EBITDA ratio of 0.6 end of year. Looking into a little more details in the segmentation. So the two operational segments, Products and Services. We saw this 27% increase in Q4, 38% for the year, and we saw a small decline in the services. But also, as Charmaine mentioned, it's not because we see the decline being something for the future. But right now, we are in this transition phase where some of the revenue is cannibalized from services into products as we deliver solutions more based on products instead of doing bespoke software development. The adjusted EBITDA for the group we saw in products, a high margin of 30.6%, which was a quite high margin. This was supported by completion ratios improving in some of the larger operation agreements where we saw in Q4 '25, just starting some new engagements where always when we start new engagements within operations, we see lower margins than when we are in the operational phase of these deliveries. In services, we also saw an increased margin, almost 20% year-over-year compared to Q4 '24. This is led by the cost reductions, better utilization and also by a lower base of employees in that extent. So overall for the quarter, we saw 18.1% margin, which was 30% increase year-over-year. For the total year, we saw an even higher increase in the product-based revenues, 38%. And as we reported in the previous quarters, there's a small decline in services, which I already talked into. The organic growth was 7%, more or less equal to the total for inorganic growth, where we saw some deconsolidation effect from November, December of securities I already talked into, but also some inorganic growth from past acquisitions. So overall, the EUR 221 million is more or less a little above the guidance range that we previously did. On the adjusted EBITDA for the group, we saw products improve overall as well with a 21% margin overall for the year, which was a 70% increase year-over-year. And we saw the services was more or less the same on EBITDA year-over-year as last year, but from a smaller cost base, a little smaller revenue, meaning that margins increased. So overall, we saw for the group total, 13.6% margin and EUR 30 million in group adjusted EBITDA. As you might notice here, we previously reported on Trifork segment adjusted EBITDA. But to make things more simple, we now report and guide, detailed here on the Trifork Group EBITDA. So all, this is within range of what was previously guided. If we look into how margins develop in products and in services. As you see in products, the orange bar or the orange curve is the quarterly profit margins, and they sometimes can be implemented or -- impacted by what I talked about before, if we are entering into new engagements or ending deliveries, et cetera. But overall, the trend line is the white line is what we will refer to, and we see a positive development over the last quarters. And this is also where we expect that to go in the future as more products and the services are delivered from our own IP to have this positive development. If you look into the services, we see a historical decline in the margins. And now we are at a turning point and have made the optimizations that we have done and talked about and also future improvements to include. So we expect this positive trend to continue in the future. Overall, the guidance for '26 is to reach EUR 230 million to EUR 240 million, meaning a growth of between 6% to 11% in organic growth. In relation to total growth, then, of course, the deconsolidation of Trifork Security have some impact since this was then to be deconsolidated from January to October. Group adjusted EBITDA, the guidance is EUR 35 million to EUR 40 million. This equals to a margin of 15% to 17%, which then would be a growth year-over-year between 16% and 32%. Now we talked into the Trifork Group and the operational segments. In our Investment Segment Labs, we look into the portfolio of the Lab companies. And here, we also saw a very positive development throughout '25. As Jorn was mentioning, we did some part exits in some investments, improving cash flow, but also proving, if you could say, the valuation of the current book value on the investments. We also now see that through several new investment rounds in our most valuable companies, we have a very solid position. So 97% of our book value is now in a strong growth path, either to be profitable or to be funded with new funds for at least the next 12 months. We here see a little different presentation of how the profits are generated and how our book value is presented for the Trifork Labs companies. So here, we see -- in the graph on the left, we see the unrealized gains and the invested cash, which then is all in all our book value. And you see the book value had decreased in '25. But then you also see on the graph to the right, the actual net realized gains, so cash in hand that we got from those investments that decreased the value in the book value. So overall, maybe a more simple way to present this that we see the composition of book value on the left hand, and we will see the actual net gains in cash from those investments on the right hand. We also published today that we're looking into the portfolio of our Lab companies and to look into how to bring even more value out of that in the future and also to look into if it makes sense to divest part of either portfolio or individual companies in order to generate even more cash to support core M&A and buybacks. Just the last touch point here into the Lab companies. So here, we're looking into the top 10 Lab companies. What you see here is the follow-up from our Capital Markets Day, where we showed more details into how the development actually is in those companies. So fundamentally, they're all on a good track and here is the average. So what you see here is that from '22 to '25, we have a doubling in relation to revenue in those companies. We see a 3x on the annual recurring revenue. And we see that in average, they have a 90% EBITDA margin. So all, we are happy with the portfolio of Lab companies that we have. And now we will move over to questions.

Thank you. That concludes the presentation part, and let's start the Q&A session. [Operator Instructions] And let's start with Poul Jessen from Danske Bank.

Yes. I have a question about products looking forward because when you report the growth you have, organic growth must be much higher than -- I think it was 27%. I calculated about 40%. Can you tell something or refresh about the revenue recognition here, how you're doing it? Is it recurring? Or is it with lumpy payments by initiation of a service, so just so that we don't extrapolate too much here in case we shouldn't?

Yes. No, I can talk into that, Poul. So thank you for the question. So we still report in the notes in Note 2.2, the details around how you could say, the revenue generation is within the products. So there, you can see how much is hardware, how much is third-party license and how much is own licenses, et cetera. So the recurring part, despite, you could say, hardware, then the remaining part is more or less the recurring part that you see in the revenue.

So it's the hardware we should take out. What about third-party licenses?

Third-party licenses would most often also be on longer agreements, several years, but just where we, you could say, deliver part of our solutions, including software licenses from some of our partners.

And they are paid as a onetime or by quarter?

If we integrate that into our solutions, where we have the operational responsibility, you could say, then it's a recurring revenue. If it's -- as we saw, you could say, in the past with Trifork Security, then some of those licenses were what we call pass-through, meaning that we actually did not deliver that as part of solution, but simply just delivering a license to a company as you can say, as a vendor. But -- so the last part is not recurring, but all the rest is recurring. So the majority is recurring.

Okay. So I assume that we are talking about domestic hosting for clients who don't want to have it placed somewhere in the world where they can control it. So if we look forward into '26, '27, I would assume that you see a very strong demand or interest in joining domestic hosting. Should we expect this to grow a lot?

I can maybe take this question.

Yes, please.

So of course, the foundation for this revenue is physical infrastructures. So it's not as easy as just prompting something in AI and then you have a data center. So there are some investments. There are some timing but when that has been said, we have been doing this for many, many years, and we are very ready for growth. And so I didn't mention before that we just opened a new data center, and we have multiples in the pipeline, and we will inform the market whenever we open a new one. And we do not take the normal building time for opening a data center. And that is a little bit our trade secret, how we do that because normally, you would expect sometimes 5 years before you have the idea and until it's fully operational and full utilized. And we can cut that time considerably down. And also because we work with partners to do this, and also, just to be a little bit advanced, we are also looking for -- and we're also looking into how these things can effectively be financed because there is also, of course, investments involved here. But to say and to answer your question, if there will be a demand, and I have not met a customer or potential customer recently who has not shown interest in talking about moving operation of data. I think there is a reaction in the market, and there is a great appetite for securing and bringing back data into the country, whether it's Switzerland, Oman, Denmark, U.K., Spain, et cetera. We will not be able to locally operate everywhere, but our ambition is to be able to do this in a handful of markets over the next, you can say, 5 to 10 years.

But if we take the guidance for '26, isn't it fair to assume that the majority of growth here then comes from products and then also from the hosting and data centers?

So we would, of course, like still to accelerate product revenue faster than service revenue. But having said that, we are not saying no to very -- you can say, service revenue that fits very well in. And there is an old saying actually that if you have a product revenue, then you can expect maybe up to 10x of that in connected service revenue. We don't necessarily see that ratio, but that's what was anticipated a few years ago. So that will follow service revenue as well, Poul. And so what you see that there was this slight decline in service, this is more an effect of reshuffling the type of service, but also customers that where we were more one-time involved with customer, and we want to be more strategic partners, as you know. So we want to, you can say, clean and refine the whole business. So either it's service on long strategic collaborations or it's product and service as well. But I would like to see service also start growing. And then you have the race between product and service. And it's too early to say who will win, but the ambition is over time that product should be up towards 50%.

The next question is about agentic AI. We have to clock AI every quarter because it goes so fast. Now it's agentic AI. How do you look at it for the Trifork business? Do you see it more as an opportunity? Or is it a risk because you can see on the share prices of companies there is a very divided risk assessment of the company. So where you look at yourself? And then also if you could add comment on how it could impact the Labs companies as they are quite small?

Yes. So -- so I want to start with the Lab companies because I think that's easy to answer. So smaller product companies that operates in a niche, I think I don't see a huge risk of them being disrupted by AI. I do see that they can grow much faster with AI. And actually, if I just want to tell about something I find like a radical story. So you know the vast amount of colleagues we have, they are data scientists, Master or PhD or Bachelors in Computer Science and all the lives that were brought up to code, okay? So in some of our boards and also the boards of our co-investors, there are actually policies not everywhere, but some where they forbid developers to develop. So what you have trained the whole life to do, you're not allowed to do that anymore. And that's a radical idea. And that's kind of made your head explode. But what the consequence of that is that we become pilots or caretakers of software generation. So we will observe all this software being generated 24/7. So if you clone yourself to become a coding agent, then there will be a lot of code generated. And the good thing is there is a need for almost endless amount of code in the world. And until now, it has just been very expensive. Now it becomes cheaper. But the coding itself is still a minor part of creating a system because you have an idea you code something, but then there's all the compliance stuff. There is the responsibility, the legal commercial aspects of all this. And every of these elements will be impacted by AI, but not necessarily just replaced. And so to get back to your first part of the question about is it a threat? So right now, I see it as a huge opportunity because we are there to help our customers to take advantage of AI, and that is a growth driver for us. You can say that once we have done that for everyone, then it's over and there's no more work. But there is -- I think we are a good -- yes, a few years in the future before that risk occurs. And we need to innovate all the time. And I just want to mention one of our latest innovation in AI. So we are one of the few companies in Europe that are actively training humanoids to do practical jobs. So where everyone is talking about white collar and agents, we are taking AI to the next level and loading humanoids with AI. And the reason why we can do that is because we have years and years of experience with vision AI. And if you think like a person, you observe the world, you map the world, you have an assignment and you go and use your vision to find something and then you do something with your hands. And that's why we are teaching robots. Now this is not a big business for us yet, but this is where we always look for the next thing.

Okay. I think we will let the next questions come through. And then if Poul has any follow-ups, he can jump back in the queue. Next question from Yiwei Zhou from SEB.

Also two questions from my side. Firstly, on this opening of the new data center, Jorn, you probably have talked a bit about it. I just want to get a confirmation that you do not expect any big change in your CapEx for the coming year, even though you are opening new data hosting capacity for the customers?

Yes. Maybe I can talk into that way. Thank you for the question, and it's very relevant. So you could say we are working with different models in relation to how to prevent that we, you could say, have a very high CapEx exposure. So we expect that to be in line with previous years. And if not, then we will communicate that, you could say, to the market as well. But we have different ways to do so that we don't have to build everything from ground up every time, which then makes our investment smaller. And in many cases, you could say it's also a collaboration between us and the customers in relation to make those CapEx investments. So we expect to be at the same level also for 2026.

Okay. Great. Very clear. And then next question is also on the AI. You mentioned that you expect this enterprise AI adoption will continue to be a driver for your business. And I was wondering if you compare to, let's say, 6 months, 9 months ago, do you see that the demand from your customers has changed? For example, if you sort of saw a change in the consulting, is it still consulting based or you are hired by the customer, start to do some AI native solutions, et cetera?

I can talk to this. Please type in Charmaine, if you have any remarks, but I will kick off. So I think it's fair to say that the AI adoption in the average company around the world today is very light, is very small. And so it doesn't really reflect the hype on Instagram or wherever you consume your news about AI and the capabilities. And so there is a big difference between what I consider reality by talking to customers and what is, you can say, the imagination of what -- how the world is. And we have been pitching these kind of AI things for almost decades. And now finally, people are interested.

We don't currently have any other questions. So I'll just give it a second. Mads Quistgaard from DNB Carnegie.

Surprised, I also have one on AI. Could you maybe talk about your dialogues with customers today? Is it still focused on cost-out projects? Or do you see like an underlying stronger interest for also doing innovation projects with AI? Because for me, Trifork, is this next-gen geeky company, which I mean in a positive way. So if we change from cost out to more focus on innovation, I think this could be a driver for you guys. So I would like your perspectives on this theme?

Yes. No, exactly. And so that's also why I actually expect to see an uptick in Consulting and Service. So the first step for customers is to kind of landscape the possibilities. Where are we right now? How can we prioritize? Where we should implement AI first? And how can we be safe that our data is not leaked? You know from the Internet right now, the biggest concern of AI is that we are -- the AI is sucking all kind of information. And companies, they are not allowed to do that. They have no interest in just letting go of all the data because for many companies, the data is actually an information, is actually the value of the business. So then they need to be sure to protect this. So prioritization of where to apply AI and making sure that it's done in a safe and protect the business value. And that I see as demand number one. And you have to bear in mind that AI is trained on public data and public data is the smallest part of the important data. All the important part of the data is behind firewalls, and I think it should stay that way. So what we help our customers with is to train AI models on private data based on the model that is trained on public. And what is a little bit scary in the world now is that AI has such a big appetite to find data to train on. So it starts making artificial data. So data that is not grounded in reality. And that's something that I think we will see in the next couple of months and quarters that, that will have some fairly negative effects on the outcome of AI if that's not stopped.

I'll wait for more questions. Yiwei from SEB.

Just a follow-up question here. Jorn, you talked about the strategy of pushing towards outcome-based pricing. I was thinking that if you can comment a bit on your contract type, if you have switched more towards a fixed price contract and from the time and material?

Yes. I think both Charmaine and Kristian, and me, we can add a little bit here. What I would start saying is that if we are going to do something on a fixed price, then we would love that we create a product together with the customer where we then have the right to sell it to other customers later. So that's the first thing I want to say. Sometimes we are pushed in to deliver a fixed price thing to a customer. And then we have to make sure that the customer is really very clear on what they want because you don't want to bid on something fixed price if it's unclear what the success criteria is. And so those type of engagement, we still like to see time and material. But there is also another element, and that is time and material is also changing because now customers clearly see that AI should be used when you produce software. So we also start pricing, you can say, software, AI when along with time and material. So you can imagine tokens being also invoiced like in the data center, we invoice energy and network traffic, and other things. So tokens and tools will be a part of the package towards the customer. So a lot of things are changing right now.

Yes. And maybe just to add a little bit additional flavor is that when we then do the fixed pricing, then it would most often be related to that we have part of the delivery based on, you could say, some of our already created IP and part of the delivery based on services where we do integration of that IP. So this product-led and driven, you could say, consultancy on top.

I just want to be 100% sure. So with the time and material contracts you are still having with your customers, are you seeing any sort of price pressure from the customer? Because one of the discussions here with investors have been that the customer will also -- yes, they also know that you are automating your process and we save a lot of hours in the coding and then they will ask for a discount. Is it something that you are seeing or if you can elaborate a bit here?

Yes. Maybe if I take a first go here. Well, as long as you could say that we deliver the services as implementation and integration, et cetera, of the products that we come in with initially, then we don't see the high price pressure. But there, we're also the experts. If we are in a -- you could say, in a battle to deliver only bespoke solutions, then we do see a price pressure on, you could say, the hourly rates in relation to deliveries. In relation to whether this is directly related to AI or not, that's a little hard to say because many customers actually, as Jorn said, the adoption of AI and you could say, being allowed actually to use AI in a larger scale in developing solutions is not really in the customers we see in Europe allowed already now. So it's only allowed in limited, you could say, space to actually do that.

We have a follow-up question from Poul Jessen from Danske Bank.

Coming into this discussion about multiyear contracts, I was just having a thought that if you had to price a contract where you do the software or installation in the first 1 or 2 years and then you have to do maintenance for the next 3 years, has it become more difficult to set a pricing on the cost 3 or 4 years out in the future? I would assume that your visibility on how the cost base will evolve over time with all the technology and AI must have become more difficult. Is that correct?

So I think it's fair to say that there's a broad consensus, and I 100% agree that the cost of software will go down. And that means that we -- when we sell a product to a customer, and I'm involved in delivering something like that just today, we will promise the customer a road map into the next 2, 3 years. And I'm fairly certain that we can deliver more than that road map because we are actually estimating that road map based on the developer speed we have today. And I'm 100% sure that we can -- there will be speed up. But we also always see when we are delivering software and products to customers that there is always you can say like a 2, 3, 5, 10x demand for more features and more stuff. So it's actually quite cool that now you can actually not disappoint. You can actually maybe finally over-deliver what was expected because software has also been extremely expensive. So when we say it's going to be cheaper, it comes from a very high cost base.

I would just add one layer on top of that, Jorn, which is what I actually like is we are now jointly on a journey of outcome-based pricing, which I think is very, very powerful. And for all parties concerned because one of the things that Trifork is very, very good at and prides itself on is its impact in that space and the value that we can create. So it's not just software provision, but actual value for our customers and organizations. So actually, I think we're seeing an increasing trend around that.

The second question, I don't know if you've touched on, I was disconnected for 5 minutes about the Lab statement you're coming. Can you give any more flavor? And I was thinking what consideration? Is it all options up on the table? And secondly, any indications of timing when you have a solution on the strategic review?

Kristian?

Yes. Thank you, Poul for the question. So you could say, overall, it is a strategic review. So we are looking into the portfolio and evaluating, you could say, all the different assets in that sense and evaluating opportunities on the different items, either as stand-alone or in a potential portfolio. So right now, there is, I mean, no horizon. We expect to conclude, you could say, internally within the first quarter. And then after that, we will -- anything equal, make a new update on where we are in that process.

But it could include that you are selling individual companies or you sell whole Labs as one and...

I mean we don't have any conclusions yet. But yes, it could be a combination. It could be individual assets, as you saw in the past. But as we stated initially here and also in relation to, you could say, the status on the Lab companies, then we have a very healthy situation right now. So that's why we believe that now is also the right time to reassess whether what makes sense, what options do we have and what makes sense also in the future. And if that can then generate, you could say, some cash flow into also doing more M&A or share buybacks, then we take that into the equation as well.

And in case that you decide to sell a 20% or 40% holding of one of the companies, do you need consent from the co-shareholders?

That could be individually in the individual companies.

But I might add, Poul, that we are very close to the other investors, the co-founders of all of our lab companies. So there is -- it's not so much a question about if we need consent. We will not do anything without seeing this as a win-win for everyone.

I think we have 2 minutes left. So Mads, if you can keep it brief and we can keep our answers brief, then please go ahead. Or maybe you forgot to take down your hand?

Also a question on Labs. I was just curious about whether -- why you see the timing right now, given all the benefits you see from AI and so on that can impact positively on the Labs business. So just guessing if this is -- the purpose is to free up cash to do more share buybacks and more M&A, I guess you down prioritize maybe a Labs business, then you potentially could acquire new companies to support the product business. But isn't it difficult to exit some of the Labs business and then invest in Trifork segment right now? Isn't that a major risk given all the certainty you at least have in your Labs business as you have today? That will be my question.

I guess, Mads, that's always -- our job is to evaluate and try to minimize the risk and optimize the profit. So you say that's the job we do there, and part of this strategic assessment. So it's a natural process for us to go through this. We do it once in a while, and now we communicate that we are looking deeper into the broader portfolio. So yes, that's my comment.

Yes, fair enough. And maybe my question is to rephrase, it is more like what is sort of the purpose from you guys in order to like do more share buybacks and then potentially do M&A. Can you talk into that?

Well, as said, it is that we are always also investigating M&A as it is currently in relation to M&A. As Charmaine, she mentioned earlier is that we focus on to see if we can have some add-on investments that could fit into the product portfolio we have in different industries. So that is what we will prioritize. But as always, you could say, if we meet a company that we find very attractable, et cetera, then we would evaluate that and potentially add that to the channel of M&A. M&A is, you could say, the priority, as the first priority.

Thank you, Mads, and thank you all for your brilliant questions. There seems to be no more time. So let's end the session here. And maybe just one quick note. Kristian, our CFO and myself, we will also be available on a Danish conference call at 1:00 hosted by HC Andersen Capital and the link you can find in the announcement we made this morning. Thank you all for your interest in Trifork, and have a good day.

Thank you, everyone.