Good morning, everybody. Welcome to Half Year Financial Report from Aiforia. My name is Jukka Tapaninen, and I'm a CEO of the company, and here with me, I have Veli-Matti Parkkonen, and he's our CFO. Hello. So today's agenda, so we are going to talk about the results in brief, a little bit about the operating environment, and then Veli-Matti will cover the financial review area, and then a little bit the strategy, how we are going forward. At the end of the session, we have a Q&A, and as this presentation is in English, but in a Q&A, of course, you can ask the questions in Finnish, and we can answer in Finnish as well. But please ask all the questions at the end of the presentation. But let's get started. So in brief, so we will go on the details on these numbers, and so we grew 25% in the first half of the year. To make a point on this area, so it's a kind of a nice growth, but okay, so it's not fully comparable to last year's first half because we changed, made some changes in our accounting system, because previously we used the Finnish accounting system, FAS, and now we are moving to more international IFRS accounting methods. So nice growth, but okay, so not exactly comparable to last year. Then the other thing, of course, which is interesting, is that, okay, so we continue to have fully international business, even though we are headquartered in Finland, but we have our customers and users around the world, so in 50 countries, over 5,000 users. Now, when we look at the numbers as well, almost 70% of the business was coming from the U.S., 25% from Europe and the rest of the world, and then about 5% from Finland. I'm not saying so that we wouldn't like to have more business from Finland, but it tells that, okay, so we are in a business, even though it's a kind of interesting niche market in healthcare, but it's really sizable market and growing fast. So we are talking about 3 billion glass slides that will be digitized and available for AI to support the diagnosis. So the market is huge, it's opening up, and it's something actually that there's an active problem recognition in a customer base, so they are really looking for solutions to solve their issues. So there's about 100,000 pathologists in the world, then cancer cases and the cases that you need to diagnose, so that's increasing rapidly. So there's a kind of a growing gap on supply and demand of the work. So and only way to sort out this gap is really to kind of use latest technologies. But going back to the numbers, so we invested quite heavily on the product development. So we have this nice platform, so we have a multiple CE-IVD mark solutions for the pathologists to use in a clinical environment. But we keep on going on investing also, that we kind of stay on the top of the curve of the technology as well. So we have about 100 people in the company, some kind of consultants included. But I guess the main point is that so when we go to numbers a little bit more deeply, so that we have a certain cost level, so we invested quite heavily after the IPO, and now we have reached probably the level that we can we can run much bigger business, but we still need to kind of keep on heavy investment on on a science side and a technology side. But, but when we start scaling up the business, so then, replicating as in a software business, then, then we can scale up and, handle much bigger volume with the similar type of, kind of resources what we have today. But numbers are interesting, so but, I think this is the biggest thing what happened in, in the first half of the year. So Mayo Clinic. Mayo is the number one hospital in the world, so there's it continuously, kind of every year, so there's this ranking, the best hospitals. There is Mayo, there's a kind of, AP-HP in Paris, so there's a kind of a Karolinska in, in, Sweden or Singapore General Hospital. But Mayo has been always in the top, and it's a kind of a benchmark in the industry. You know, we've been talking about this now about two years on the same case, but it's really important. So because we made three deals with Mayo. One was a joint development, so Mayo's pathologists are using our technology. They are building their own AI models, and we have an agreement, a model, how we can start commercializing some of these models. So that's really important. And what has been already about a year in the production is the research side. So Mayo's pathologists, plus 70 people, so they are using our technology, building their own AI models, and speeding up their research work. So that's really important. But probably the most important, from our strategy execution point of view, is the clinical side. At the end of Q1, we got live the first AI model that takes care of the patient samples. And as you know, Mayo has 2.6 million slides that they are handling every year. We started small, so we have a Ki-67, which is a kind of breast cancer biomarker. But that's on production, and it's set up with the Google Cloud, Leica scanners, and Sectra PACS, picture management system. And whenever there's a kind of slide that has a Ki-67, so then it goes to Aiforia to be diagnosed, and we do the quantitative analysis, give the report, and start automating the whole workflow. As far as I know, so this is the kind of biggest and maybe the only real kind of live implementation of the CE-IVD-marked model in the world. And it's a huge reference for us. In the U.S., by the way, so they use what they call kind of a lab-developed test, so that they kind of certify themselves what they can use. So the FDA is not yet an issue because they can add more AI models and certify by themselves and put it in a work. But the point is that, okay, so we are giving the kind of support for the diagnostic, and then the pathologist is always kind of verifying the results, but we give a huge kind of a benefit on a time saving, accuracy, and so on. So this is a big thing for us because we can use that as a reference for the other customers. And of course, we have multiple other cases going on in the U.S., but that's Mayo, number one. But two other things that I would like to mention about the first half of the year. So in the European area, Veneto. So Veneto is a region in northern Italy, and they got the huge funding from the EU. Thanks, EU, for that. So for the digitization of the healthcare systems in the hospitals, actually, they were talking about plus 100 million, and they reserved about 10 million for pathology. So and that contains the scanner acquisition, so laboratory systems, and then the AI. And we are taking care of the AI part. The first part of the deal was EUR 1.2 million, three-year deal, 12 hospitals, 5 AI models, totally 200,000 samples. And this is really meaningful, but now, now we really kind of have a good case in Europe as well, in a clinical site, to use as a reference in the future. And the first deliveries of this will happen in the second half of the year. If you think about the numbers, so, the EUR 1.2 million, we didn't book any revenue out of that EUR 1.2 million in the first half of the year. Some of that revenue can be booked in the second half. Meaning that, okay, so we start delivering and then we recognize the revenue. Then U.K. Now, as we separate the U.K. from the European Union, it's important to have a kind of a solid reference in the U.K. as well. And, we earlier this year, so we won a deal with the NHS. So they were validating and evaluating different AI systems, and we qualified on a prostate and lung cancer. In a prostate, so there's kind of two other kind of vendors, and in the lung, that's perfect because we are the only one. If the NHS is buying an AI model for lung, so then they buy from Aiforia. So but that was covering 25 hospitals, so we actually did the first closing on that area. We haven't said before, but it's a Southampton Hospital what we won. And that is also something that, okay, so deal was done in the first half of the year, and revenue recognition most likely happen in the second half of the year. So that, okay, not seeing it in the revenue numbers, but really really happy of these results so that we are progressing in all fronts. US, solid case in a production in Europe, huge, huge kind of a case that gives us a referenceable customer and the same in the UK. But moving forward on that. And of course, we have the platform that the pathologist can build their own AI models. That is a core of our offering, and then we use the same technology to build those ready-made AI models for the clinical side. But we haven't forgotten the pharma market, so there's a-- there's already a lot of customers, and we won couple of additional ones in the first half of the year. One was Orion, so we gave that name out. The other one is the one of the, let's say, top five pharma companies, but we were not allowed to publish the name, so but that, that's a significant deal from, let's say, Switzerland. Maybe I don't say the city, so then you guess the company name, but but we are progressing on that front as well. So and we can offer them nice tools to speed up the drug development processes. So that okay, because they do a lot of animal testing and a lot of samples that they need to analyze in in that area. So and with the AI, they can do it faster, take the clear cases out, focus on the ones that actually are meaningful, and improve their processes on that front. On the funding side, so, we actually got a loan commitment from Business Finland. We are very thankful that we got a kind of a loan commitment. We haven't used it yet. We actually used it; we took EUR 2 million out of it, roughly, and we have EUR 5 million left, so, that's something that gives us a buffer if needed. But okay, so that's a very, very kind of a good thing for us. The operating environment. So I mentioned a little bit so that the market is there. So and as I said, so there, there's two things. So we are not in a market so that we are inventing something new. So this is something old process that we are now automating with the modern technology. So as there is a huge demand and a gap on the resources and what can be taken care of, so the demand is there, and the customers are actively looking for solutions, and we can provide that. And but we are not the only one. So this is the—this picture is actually kind of showing the environment that we are in, because there's always this physical sample that will be digitized, then the image will be stored and distributed to pathologists, and all this kind of a flow of things that needs to happen. And then, of course, linking all the systems to laboratory systems or hospital systems. But our position in this workflow is the critical one, because we are the one who actually understands what's on that slide. We do the quantitative analysis. We took all the cell counting and the distances, whatever you need to find from the slide. We populate the report automatically, giving all the things that the pathology needs to kind of put in his or her report. And then the pathologist is kind of confirming whether he or she agrees with the AI. So in a way, we think that that's the place where the value is created. And then when the report is ready, so it goes to a doctor who takes care of the treatment decision, and then that completes the flow. My point is that, okay, so we've been now kind of expanding our partnerships and ecosystem in a way so that we work with all the kind of big cloud providers, whether it's Amazon or Microsoft or Google. They actually cover about 90% of the world's traffic. So we have a kind of system working with all of those. We have built the interfaces, integration to kind of all the surrounding systems so that we can actually deliver. Because our thinking is that, okay, when you talk about the AI, it's not just making a kind of one algorithm. So it's something that you build a solution, and you bring that solution to a kind of environment, so that you need to be able to deliver, and you need to be able to overanalyze that. So because the customer needs to get return on their investment, so and the investment is coming, so when they get kind of... You can do more things on the same time, or you could do it faster, or you could do it more accurate way, and, and then you have a full portfolio of tools that covers most of the disease areas. So but, that's the environment. It's now kind of rapidly expanding, so we have now the first kind of a big customers, and the position is quite nice. But I think I was talking already too much. I will hand over to Veli-Matti to talk about the numbers, and, then we go back to the future and, and the questions. Ah, you have one. I have one. Thank you. Thank you, Jukka. We have a pretty busy six months behind us, and, the Mayo Clinic go live, this period was amazing, achievement for us. I have to repeat it. Think about it so that, our application is used to save lives or helping to save lives somewhere in the world at the moment. Here. Figures. As Jukka mentioned, that, we changed our revenue recognition practices to comply with, IFRS systems, IFRS standards, this report. The international financial reporting standards, they make it easier for investors to compare companies cross-border, while as, national, accounting standards, leave more room for interpretations. I have to admit, we still report with a, FAS, which is the Finnish accounting standard. The change made, some, 25% of our revenue, comparable revenue, pushed, forward. The biggest changes were made in projects, how we recognize projects. We used to recognize them like the completion percentage, but now we recognize them only when they are completed. The projects don't pose a big issue for us going forward, but the way now, when we are shifting our sales from the research sales to tenfold bigger clinical deals, and those deals always include some kind of integration project. Like Veneto and NHS that Jukka mentioned, we haven't booked any revenue from those yet. The revenue we reported for this was EUR 920, showing growth of 25%. Without the changes, we would have landed at a decent 70% growth level. We added on this report a new key figure, the order book, which gives more color on our progress going on. EUR 2.36 million is over 100% more than from the year end and over 200% more than a year ago. The EBITDA was a negative EUR 4.7 million. EBIT was at negative EUR 6.2 million, and the net loss for the period was EUR 6.3 million negative. Equity ratio stayed at a good level at 78%, and we had cash EUR 19 million. That plus the EUR 5 million of undrawn credit facility from Business Finland makes and gives us a good buffer going forward. We had on average 69 persons, and the personal expenses were EUR 3.8 million. We invested EUR 3.1 million or made investments for EUR 3.1 million, and the EPS landed at -0.24 EUR. We still invest heavily in R&D. Over 30% of our cost base goes to that direction. On the other hand, our fixed costs are stabilizing. We still recruit new personnel, but trim on other parts of the organization where we see that resources are not needed that much. While the costs are stabilizing and the order book growth, forecast growth for the top line, that gradually pulls us towards the black figures. Back in 2021, when we did the IPO, we had an internal forecast that the pivot point for our result was the year 2023. Cash flow from operations, that shows that our business model is asset light. We don't need capital to maintain our growth or the business. The working capital delta is zero or even negative. So, we are happy with this, but of course, we need some monies for the future growth since we are not there on a product-wise yet. But anyhow, we maintain our promise to be cash flow positive on operations by the end of 2025. Thanks. Okay, so let's continue about with the growth strategy. So these were the short-term targets that we announced when we did the IPO almost two years ago. Getting the CE-IVD mark portfolio up and running, so we have done five. So we actually have a sixth, which is a viewer. And, but then, of course, in Europe, so they changed the regulation a bit, or at least the process, how to kind of certify these things. So there was a little bit kind of a time gap when they didn't have a kind of a notifying body, so who could certify things. But now we believe that before end of the year, which was actually our target, we will have the sixth model in play. Of course, we have a, we have a multiple models done with the, with the technology. So there's plus 400 models done, so there's 80, 80 + kind of scientific publications and the posters done, so those type of things. But okay, so we want to kind of move forward with the CE-IVD mark, deliver what we promise, and then, then we expand this portfolio rapidly when, when we get the customer demand. So but now the first clinical cases we can handle with the, with the current portfolio. But, but as, as I said before, the strategy is to, to have a good coverage of, different kind of disease areas, and, and that requires multiple, multiple AI models. Five customers in the clinical side, yes, done. The large pharmaceutical customers, done. Over 5,000 users, done. So these were the goals till end of 2023. So we have done it, so I guess we need to kind of update for the next event a little bit what the targets going forwards are. But we are happy from that perspective, so that what we promised, so we delivered. So but then this is, of course, maybe more interesting now to look at as the market is opening up. So we have the solution, we have the first references, and the adaptation is kind of speeding up. The customers are kind of taking on the technology and start using it. So that we still need to expand the offering, what we have, but we might also kind of give a little bit in depth, so that we pick and choose certain areas that we go deeper and see what else can be done. Because with the technology, there is a next level after the... Just the kind of speeding up and giving the accuracy, the whole efficiency thing. So but then, then you can start building a new type of things, like Mayo Clinic has done, model for QuantCRC, where they kind of use our AI to run the analysis using a patient data and combining those two, and based on those two things, so they can predict what is the patient outcome. And they use that tool to decide whether they give a three-month chemo or six-month chemo. But the digitization and AI opens up new type of opportunities, and obviously we are looking very carefully, so where the market is going on, on that side. Veli-Matti mentioned the happy, happy 25 when we start making, end of the year, making more money than investing. And, as said, so we are on, on track on doing that. And the revenues, by the end of the decade, so we should be a +EUR 100 million company for sure. So that, if all, all, all of these assumptions are true, so that, okay, there's a handful of players actually in this huge market who can deliver something. We feel that, okay, so we are actually best positioned from a technology perspective, but of course, understanding so that, that there's a competition out there, which is actually a good thing because it means that the kind of certifications for this type of a solution will happen. So the customers are adapting the technology, and we are competing which one is good and which one is okay. And so far, we've been performing very well when we are kind of in a head-to-head comparison to any of these kind of competing companies. User base needs to grow, obviously, and it will grow when we sell more. Like with the Mayo, so I always jump back to that topic. I mentioned that there's 70 people in the research side using our tools to make their AI models, but there's overall over 300 people in the Mayo Clinic in Rochester, somewhat kind of a link to Aiforia solution. That gives certain kind of a comfortable position for us, so that it's a big investment from a customer side, and there's a certain stickiness when you kind of get the system up and running, so and you can start adding more AI models and customers are staying. Key accounts, big accounts, obviously, so we are kind of heavily focusing on kind of building a pipeline, closing more deals. And now we are moving because so far we've been always doing demos, very tailored demos with all of these hospitals, because you need to show that, okay, it really works. But we can, we can actually speed up the sales cycle a little bit because we have start having a references. So and we can bring a best, best practice from a Mayo, saying that, okay, they are using these solutions. This is where you implement. These are the security things you need to take care of. And we have, we have learned a lot from that process, and we can start replicating what we have done before. So every new implementation is, is quicker and faster and, and giving a better ROI for the customer as well. 2023 and beyond. So obviously, so as I said, building up the solution portfolio and then adapting new regulatory frameworks both in EU and in the U.S. I know that in every single event, somebody is asking, so whether we are doing the FDA and when, and is it needed? As I mentioned before, so now we can, we can sell to these big hospitals, because they use the LDT, laboratory developed test method. But obviously, so we have started to kind of look at what the FDA requirements are, how the process goes, and started to take the first steps. But I don't want to give any timeline when it's happening. So but, at the moment, it's not blocking any sales. There are some other regulation in the U.S. as well, so that we've been already tackling, and we come out with some more detailed information shortly. But so we are following up how the environment on this area is changing both in key markets in the U.S. and Europe. But so far, we have all the kind of certifications and things what is needed to sell successfully, both in kind of a preclinical and clinical side. Continue developing a solution, of course, and then we want to focus. We have some kind of good customers in Asia as well, and we take care of those. But really the kind of a limited sales team, what we have, we are focusing on Europe and focusing on the U.S. because those are the kind of biggest market at this point of a time. So I read from some studies so that in the world, so there's 102,000 pathologists, and was it the 90% are in the 10 countries in the world. So gives us a little bit the indication, so where the potential users are. But obviously, so we look at the whole kind of big play. But okay, so we need to start somewhere and focus on somewhere getting the results. Then the other thing that we are already developing and going to develop further is the ecosystem play, because I mentioned the technology partnerships like the Googles of the world and Microsoft of the world. So they are kind of a given because we are cloud native and are running on a cloud. But then there's a plenty of other things that, okay, so we need to partner with. And maybe it's this kind of a AI, kind of a hype that is going on, so but a lot of big names, they all kind of approach us and want to work with us, which is very nice, and we are happy to do so, but okay, so we can complement the offerings what they have. But then on the selling side as well, so we need to at some point, kind of start ramping up and getting more feet on the street, using, using the channel. We have Epredia, which is working nicely, but we, we, also want to expand and, having more coverage. And that means that, okay, there are partners who can, let's say, kind of, open up some doors and, and then, then we run the selling. And in the future, we probably have some partners that we kind of, train and certify, and, they can, they can take care of part of the sales cycle and part of the delivery. So that's way to scale up. But we are also very realistic, so that, okay, so when you are in an early market, t he vendor has to close the deals, and the vendor has to deliver the first deals. When you start having the references, the channel will get more interested in... And then, of course, expanding these accounts. We have a good number of customers already, so we can sell more. The thing is that, okay, first you implement a couple of AI models, but there's plenty of room to expand those accounts and a footprint in those customers. But that's it. Ready to take some questions, and hopefully giving some answers as well. Good. Hi, it's Antti Luiro from Inderes. One question first on the order book and what it consists of. So it's around EUR 2.4 million. How recurring is this order book in nature? What is the kind of rough timeline when this would be booked as revenue? Any other comments on what's inside this figure? Okay, first, we haven't disclosed too much of that, but anyhow, I mean, obviously, I mean, the order book includes those two deals that we have mentioned in this session, so, Veneto and NHS. So those are our three-year deals. But other than that, it's pretty much max 12-month deals. So, invoiced, already invoiced deals that we have started to book something, but since they are license based deals, we book them as a monthly recurring income for us. Right. Maybe also to point out how we sell, so the order book, as I understood it, is kind of a remaining performance obligation, so that we have already contractually kind of a guarantee, so that in case they need to pay. But then we have a lot of customers that, okay, so we are renewing the licenses, so those are not part of the order book, so- No, no, no, no. Of course, no. Yeah. I mean- Yeah. ... all of them are kind of like, books that are... or orders that are paid to us, or they have purchase order for that amount of money. And, some of those are already paid for us, invoiced, and, they are, I would say, max 12 months, orders still running there. And, we have recognized them as monthly revenue for us. But now when we are shifting over to more clinical deals, normally these clients want to do more-than-one-year deals because it requires quite a lot of effort from their side as well, the implementation and checking the security systems, and so forth. So, you could see those on the Veneto and NHS. On the other hand, Mayo has... I mean, those deals are only, like, one year from now on, and we just continue those. Mm. Right. Hopefully. So maybe not worth it to overanalyze it. Yeah. ... at this point, since the form of the business- Yeah. ... is still changing. But it's some indication of what's coming, you know, roughly a year forward, and of course, most of your business is something that the customers keep doing, right? So in that sense, there is a certain recurring nature over a period of time. Absolutely. Yeah. Absolutely. To simplify, so that, it's maximum three years—... right? Yeah, that's absolutely. Some of these contracts have a kind of a option to renew- Mm. ... with a similar time. But actually, we do have... I mean, by that I mean that only those two are more than one year. Yeah. All right. Got it. That's really helpful. Okay. Another one on revenue recognition, a little bit on what happened in H1. You mentioned that what happened with the IFRS change in revenue recognition was affecting mostly projects and how they're booked. And geographically, you had a drop, for example, in Finland. I know absolute terms, it's a small, small number, but was this also due to projects being scaled down but not kind of losing users from your software? I think the biggest part what happened in Finland was that some of the credits that we sell, they were booked as how much are consumed, and we kind of like, we had to go down with the amount of money that we have recognized as revenue. Right. That's perhaps the best answer for the Finnish side. Right. Right. So this speaks sort of to the level of recurring nature of your revenue. It's not exactly month-by-month recurring, but rather recurring, so that- Yeah, I mean, for the, I mean, analysis that now we have one client, actually, it's Mayo, on live, and then the credits, we are booking them now, on a consumption basis. So we used to do that, okay, you have a contract for the license for one year, and then even the credits for the clients, we book them, like, 12 times the amount what they paid for the credits. And now we calculated how much they have used the credits, and it's the exact amount that we have delivered- Yeah. ... the service for the client. Yeah. And- So and, keep in mind, so that most of the existing or older customers are research and universities and some pharma companies, so that they are consumption-based and so on. So it's kind of a different... When we go to a clinical side, then the deal size is much bigger, but okay, so then we have multi-year contracts and secured revenue, so. Right. Right. So there's, simply put, some cyclicality in the level of use- Mm. ... in the existing customers, but no kind of customer losses that would be- No, no. No. Yeah. Exactly. Yeah, that's helpful. On the CE-IVD models, you mentioned that the notified bodies have not been operating for some time, but now, in H2, maybe you're able to put in some models- Right, ... through that process. How is that bottleneck looking? Is there a lot of kind of models and work for these notified bodies that is kind of stuck up there and not moving forward? And do you believe that going forward, the flow is gonna, you know- Yeah. So I think they- Get better phase. ... Yeah, it took, because they changed the regulation, I think it was last year, on a May timeframe. So and then, first there was nobody to certify anybody. So and then, and after that, I understood so that there's a kind of a huge pressure to kind of, certify a lot of things, and there was a bigger companies, bigger players, who took a lot of bandwidth from those companies, so that it was not easy to find a place to do the certification. Now we are kind of, we have found some, and then we will do some, but I assume so that it will start kind of, working now faster. And it's called IVDR, from going forward. All right. Thank you. Also on your sales work and the split between your direct sales and partner-sourced deals, you mentioned you wanna go deeper into the partner side- Mm-hmm, ... and get more partners activate there, but how has it been for now between partners and your own sourced deals? Yeah, so we did some business with Epredia, so that's they brought us customers, they closed the deals, and but it's quite a small amount. Most of the deals are something so that we collaborate with either Epredia, or we collaborate with Leica or 3DHISTECH, one of the scanner manufacturers, and there's a kind of a tender out. So but we have the same goal, because we need to close the deal, we need to work together, so that type of a partnering or then the big players like the Roche and Hologic, and these guys, they want to kind of expand their own offerings. So but it's more kind of having the same goal. And Sectra is a good example as well. So it's a very successful company from Sweden, and they are in Mayo, they have many, many French customers and all over the world, so they're quite successful. So we have built the integration with them, so now because they can't really sell their solution without an AI, and we can't sell our solution without them. So it's kind of... Even though we don't do any kind of commission payments or revenue share or something like that, so both will close their own deals, but we need to exist in a deal at the same time. Yeah. Yeah. So maybe the benefits come from deal sourcing, but then you also need to put in the work, of course. Exactly. Yeah. Exactly. All right. Then finally, on the bottlenecks of your revenue growth, I know this is a tough question because you have so many different customer segments and different phases there, but, you know, getting to the positive cash flow, situation end of 2025, of course, takes growth. And now looking at H1, of course, you've won many deals, but the revenue growth was smaller. If we think of the bottlenecks right now, what are the biggest things slowing down your revenue growth, if we look forward the next year or two? Well, that's a kind of don't want to go too much on the negatives on this, but okay, so it's of course a fact. So it's a kind of a new market, even though there's a demand and it's opening up. With our scale, so we can deliver, and we can make what we are now promising to make, but to kind of make it big, it would mean that there has to be something happening on a customer side as well. So because in the U.S., there's reimbursement codes, there's for the breast cancer existing, they can use the AI, so that helps on selling. In Europe, so there's a lot of a public sector, and public sector is struggling with the, with the budgets and that type of thing, so that somebody needs to see the bigger picture so that, okay, automation is the way to kind of sort out some issues. Like in Finland, so there was some news that they changed the kind of how they pay the people, and then the system was kind of getting kind of in a big problem. But okay, so you need to use the technology to help out or something else to sort out the core of the issue, and then you fix the other things. But somebody needs to kind of see the big picture, provide the budgets. As we see in the Italian case, so there was, I think there was, four integrators who were kind of getting the money from the European Union and then providing all the services and the solutions to that region. So but somebody was taking an initiative to getting a budget. But if you think about the buying process, so that each one of these hospitals would start kind of figuring out and selling internally and finding a budget from something else, so that slowing down. So somebody needs to put the kind of a big thing rolling, so that's the one thing, and giving the money and seeing that, okay, so it needs to be done, and that the benefits are obvious. The second thing is the regulation, so more standardization is needed, so whether it's a CE-IVD or FDA, but there has to be kind of for the kind of image quality certain standards, so that it's easier to build an AI models. Now, we can work, of course, with the multiple scanners and file formats and so on, but if that would standardize, would be much easier to kind of build on a scalable AI models. And there are a lot of things that needs to happen. From our side, so we are looking what's going on in the market, so I'm trying to adapt. But those type of things would help us on growing faster as well. Right. So if I take this correctly, the bottlenecks maybe are more on the kind of just demand picking up in a new industry, but when the customers are won, they're moving nicely forward. So it's mostly just getting those deals in than the movement in the market. Yeah. So and, yeah, we closed a smaller deal in, for example, in Paris, with the AP-HP, Assistance Publique-Hôpitaux de Paris, so which is a huge organization. They and they have a lot of money. They I think they have 15,000 doctors and, yeah thirty-40 hospitals in just in one city. So it's a huge organization, but they are also kind of struggling to figure out the budget, or then they have some concerns working with different cloud providers. So there are a lot of things that that needs to be kind of just clarified because the security is there. Cloud is probably much more secure than having an own kind of servers running in a hospital. Those type of things are slowing down, but the budget is one big thing so that we need to sort out. All right. That's very helpful. That's all for me. Thanks. We have few questions from online as well. First, regarding Mayo Collaboration, can you still clarify as for clinical implementation in the Mayo Collaboration, do you expect this to happen in H1 or H2 in 2024? On a clinical side, so it's already in production. So that, that happened already kind of, three months, four months ago, and we have one breast cancer AI model in a live production. And then our goal is to add one or two models before end of the year. So, and the end goal is having, of course, much more, but okay, so this year, this year in a fiscal year 2023, so we are going to add one or two more additional models. So has Mayo committed to implement any other of your solutions into their clinical workflow beyond the colorectal cancer solution? Yeah, the colorectal, yeah, that was a good kind of a pick up what I said. So colorectal cancer is one of those jointly developed models, and so they have done it. And now we are in a discussion, so that's probably the next one. So they're going to implement from their own kind of a built models. And then, of course, we add something from our side as well. Thank you. Going back to the order book, in one of your previous slides, I think I saw the order value for Veneto was EUR 1.2 million. So is it correct to assume that approximately half of your current order book is related to this one project, since you did not book anything from Veneto in H1? That's correct. Thank you. When do you expect all of the projects in your order book will be finalized? Well, within the next three years, that's the maximum age or time span for any of those contracts. Yeah, I can open up the Veneto a little bit more. So it's a three-year contract, EUR 1.2 million. There's a certain upside because there's a little bit... It's a-- it was a public tender, so there's a little bit upside available as well. But the current setup is divided roughly to three years. In the first year, there might be a little bit more coming from a implementation work, but then, then we going forward, so then it's a kind of a steady, if I say, +EUR 400,000 a year. Wonderful. Do you have any potential customers in Spain? Yes, we do. That's great. Sectra Amplifier Marketplace has 36 AI models, including Aiforia ones. What do you think will be role of this kind of marketplaces in the future, and how important are they in your plans? Well, so we will participate. So we have maybe a... It's kind of a we'll, we'll need to see how it will plays out. There are some players who thinks that, okay, there will be companies building an AI models, and somebody will take care of the kind of a user interface. But my background is from a software industry, and I have seen very few cases that it has really worked well. Because if you start implementing a kind of a solution in a clinical environment, taking care of all the security, all the kind of a different connections, so and if you start having a kind of a different vendor for different AI model, so you need to be implement and maintain all of these different solutions. So I believe that, eventually kind of a platform approach with the vendor being able to provide multiple models will be the winning solution. But, we are happy to work with Sectra. We are happy to be in their marketplace and some others as well, but, let's see how that plays out. So we are not counting too much revenue on that. So at this phase of the market, so I think it's something that we need to sell, and we are selling together with Sectra, but it's a kind of an enterprise type of a sales process, especially for the clinical. Thank you. And to conclude, can you provide an update on your run rate? How far will your current cash position take you? Maybe Veli-Matti takes that. Yeah, I mean, as I've told you that we are aiming to be cash flow positive on operations by the end of 2025, so everybody could calculate. Of course, then you have some cost for the investments, but pretty soon after that, we would be cash flow positive altogether. So, we are happy currently with our cash position, and we monitor it very, very closely, so that that would we need more money, we would act, but at the moment, we are happy with this. Yeah. So I think I would see in a situation where we raise more money when we see that the market is opening up quicker than we think. With the current speed, so we can handle, but, okay, so if something happens and there's a new opportunity, so we might want to invest more. And as I said before, the cost level what we have now probably stays in a ballpark on that level, meaning that, okay, so when we keep on growing the sales, so then, then of course, the finance or the kind of a profit situation gets better. So, and, currently we are planning on making those things what we are planning to do. And we've been making some nice kind of a savings on a cloud costs and in the other areas, kind of simplifying things. But okay, so that's maybe another other story. And maybe one to two kind of cost-related issues is the access to your data, because that has been always kind of a top question. So do we have access to your data? Some competitors have a kind of big data collections. What I can say, so that we have a technology, so that okay requires a little bit less slides for training, so that okay so we don't need so much material, but we need get the material from the customers, which is an important thing, so that we kind of modify and configure for their needs. But more most important thing is that okay so now we have access to a kind of Mayo's data collection, which is the biggest in the world. And it's not just Mayo's. There's Mount Sinai slides and a couple of other hospitals, but that, that's the biggest collection in the world. So that, that's something that the Mayo collaboration is opening up for us, and just wanted to point out. So that's a cost-related issue as well because we get an access. There's actually one more Mayo-related question coming in. What is the value from the AI model from your point of view in Mayo, whether it is they're developed, whether it is developed by them or by us or by Aiforia? Well, I think the quality-wise, so it has to perform in the same level anyhow, so that whether it comes from our science team or their science team, from our perspective, of course, it gives us more resources to get the more AI models faster. And from a salesperson's perspective, so I think it's quite nice if we can start commercializing some of these models together, mentioning so that this is a quality standard from Mayo, so maybe that helps to close deals faster. But that we will see when we start kind of commercializing those. Thank you. No more questions in online. Okay. Thank you so much. Thank you.