Good morning, and welcome to Aiforia's earnings announcement today. We are going to talk about 2020 and a little bit about the 2023 as well. With me I have here Veli-Matti Parkkonen, our CFO, and my name is Jukka Tapaninen, and I'm the CEO of Aiforia. Let's get started. 2022 was an interesting year for us, it was really kind of a first year after the IPO and we got nice investments in, back in December 2021 and we gave some guidance and ideas what we are going to do in this year, how we are going to invest on R&D and product development and customer acquisition as well.
We started to move forward with that plan and I'm very pleased so that the things that we were planning to do in 2022, we actually we were able to achieve. On a revenue side, we grew roughly a little bit less than 100% in the annual level and the sales was actually accelerating in the second half of the year, so second half of the year was 107% growth comparing to previous similar half of year session. We are very pleased on that, so we gain nice new customers, and I will come back to those topics a little bit later in the presentation.
The other side, of course, is the investments. As we said, when we raised money, we were planning to invest on a growth and there was two main things. One was the R&D and the other one was the customer access and the employees, what we needed for that. We invested EUR 5.9 million on our product development last year, so significant growth from the previous year and we also increased the number of employees, so we have 65.
Overall in the team we have a little bit over 100 people today. We actually strengthen quite a bit the R&D side, both on a science and technology and then we're building kind of a sales engine in the U.S.. We have a U.S. office in Boston and now we are opening up a office in the Rochester close to the Mayo Clinic as well. What comes to geographical distribution of the revenue, the U.S. business was growing nicely, 253%, but we are as you all know, we are still talking about the small numbers. Couple of deals will make a big difference.
We are very pleased on the U.S. growth and Finland was 14%, Europe and rest of the world 18%, all in all, nice progress on the sales side. I think the key thing to think about here is that, okay, we were in a multiple competitions, competitive situations, typically we came out as a winner. That gives us a good proof point that the solution is well accepted and competitive in a marketplace. Veli-Matti goes on the numbers more on the details a little bit later. Couple of names, Mayo Clinic, this is actually the biggest deal what we have so far and that was a major competition that we won already in 2021.
2022 was time to deliver on research side we actually made three contracts with Mayo. One is the research side. There's over 100 pathologists using our platform and building their own AI models. The second thing is that we sold AI models that they will take live on a production on a clinical environment. In Mayo, there's a 2.6 million slides, it's big opportunity for us and we've been working on that topic.
Third agreement what we made with Mayo is the joint development agreement, which is kind of will be the next step that we start executing. But two first ones now delivered and on Q1, so we should have first AI model live, breast cancer Ki-67 model. I think we've been talking about Mayo in every single event, but it's so important for us because we believe that that's the best practice for the whole industry. There's an ecosystem there, other vendors as well, but the solution comes out from Aiforia doing the diagnostic support, bringing the AI models, bringing the platform. But there's Sectra, there's Leica, and all running on a Google Cloud environment
That's the combination so that we believe that will be best practice for the industry so and the other big organizations will start following up. Of course, thanks to Mayo and thanks to our sales efforts, we won a couple of nice other accounts as well on a clinical side. One is City of Hope. That's a well-known hospital in U.S. in Southern California. Actually, it's a 30 hospital. Now we sold the proof of concept solution.
Proof of concept is done, so now we go to the next stage with the City of Hope going forward, so that's one of the big opportunities to have a revenue impact and also a nice reference customer. Wake Forest, big organization, we sold to their research side and now the next step is to expand on a clinical side as well. Just putting in a perspective with the Wake Forest, there's about 90 hospitals. Later on I will talk about the NHS. We made a nice progress there as well. These U.S. organizations, they are extremely big.
We can land and expand with those accounts so that, okay, when we do the first level of implementation, then we start adding the AI models as we have in our strategy. There was a lot of other things happening as well, of course. When we did the IPO, we promised that we will deliver new CE-IVD mark solutions. We actually delivered four AI models and also clinical viewer, CE-IVD marked. Also we got two U.S. patents, and we work a lot on security topics as well. We get nice certifications, ISO 27001, SOC 2 for the U.S.
This was actually coming from mostly from a Mayo requirements, but it was a good exercise for us also to improve those type of things, because going forward, I believe that all the new customers will require exactly the same things. It was a good work that we did there. On the marketing and communication side, we launched an Aiforia Community Platform, and that's a a collaboration space, so that okay, the industry gets together and share tips and tricks and also possible to share some annotations and algorithms.
Our idea was to be the kind of a de facto, a leader in the industry, building this type of a community site that the pathologists can get together and share ideas. Of course, on a team side, we strengthened the sales teams and, as I said, R&D and science team as well. We also strengthened a little bit the management team as well. Tom Westerling moved from a other position to be the CCO of the U.S. operations. That was a very, very strong add-on on the team as well. The Epredia partnership is not a new thing, it was actually coming from 2021 as well.
It took some time to train the team and make it operational there. Now we start seeing the results. Wake Forest was one of those accounts in the U.S. that we did together with Epredia, and now we see that there will be some movement in the European accounts as well. Epredia is one partner, but we also look at the other opportunities because this is an interesting ecosystem. Partnering is a way to scale up quickly on a sales side as well. Good. The operating environment. We've been talking about the market and when the market is opening up. There was some kind of...
I think, some of our partner was saying that, when the market is opening up, it will happen a little bit slow, but okay, when it happens, it's actually happened bigger than anybody expects that. What we see today, there's about 100,000 pathologists in the world. There's about 3 billion glass slides that needs to be digitized. Digitization, percentage has been growing from a 15% going to a 37% in the next coming years. There's enough digitized slides in the market. The main thing is that the population is aging, so and unfortunately cancer cases are going up. There's a growing demand on the other side, which is happening and will continue to happen.
On the other side, there's a certain amount of resources and old traditional ways of doing things. The automation is the way, and AI is the way to improve the situation and close the gap. There, there is no question at all if the market is there. We believe that, okay, the kind of a fast adaptation and the market opening is happening right now. This is an old slide, but okay, just giving an idea what we do. Moving from a 150-year-old process of manually looking at those tissue samples on glass slides, going to AI-assisted diagnostic. There's a very important role for the pathologist in this play.
Our technology, AI, support them with the quantitative analysis, giving all the calculations, pre-populated report from that slide, and also showing in a screen where the possible problem areas are. What we can provide in this process is, it's more accuracy and, in a some level also things that the human eye cannot see, but also huge amount of efficiency on this process, time savings and faster diagnosis, more accurate diagnosis, leading to a better patient outcomes. First of all, right treatment decisions and then better patient outcomes as well. This is the process.
I'm not going to go too much on a detail, but that's why I said, there's another players in this flow as well. We are not doing the wet lab part, preparing the samples in the glass plates and doing the stainings. There are scanners, there's PACS and image management systems, moving the slides. Then of course, the LIS systems, the laboratory and hospital systems that we need to interact with. We believe that okay, we are actually very nicely positioned in this flow because the value creation is happening there where we are, that we give the diagnostic support and all of those calculations that I mentioned.
That's because at the end of the day, on this whole process, what matters is actually how do you diagnose and what type of a treatment decision you make at the end of the day. Those are the important things. You need to have, of course, all the other components as well and that's why I said partnering is a key thing, and we have APIs to kind of surrounding systems and a cooperation going on with the multiple companies on this field. Well, we believe that we have the best product in the market, and value-adding product rightly positioned in this workflow. Of course, how do we think that that's the case?
If we look at the competition, so somebody using all the machine learning-based technology. We use the deep learning-based technology. Some companies are just analyzing the slides, saying that, "Okay, this might have a cancer and this doesn't have a cancer," and then the pathology need to do the work as before. In our case, we can go a step deeper, so we do actually kind of a cell counting and a distances, and we give the information that the pathology actually needs for the decision making. An outcome is that, okay, so it's faster and more accurate, but the proof point on this is what I said earlier, so that in most of the competitive situations with the competition, so we win.
That leads to a customer. I've been showing many times the logos of the customers. We have the best ones in the industry as customers, either using a proof of concept or using in a research setting or using it now like in Mayo, going to a clinical side and using in a clinical production. We have the Mayos and AstraZeneca's and Novartis. We have our customers from a clinical side and a pharma side.
Because in pharma, we've been talking a lot about the clinical, but the pharma is as important because we can bring a lot of value on shortening the drug development cycles, shorter time to market and these big companies can make a huge cost savings using our technology. As said, Novartis, AstraZeneca, BMS, multiple other companies using our technology, and then of course the university and research and that side of the business as well. Enough proof points. On a money side, people side, skill side, we have a good amount of a growth capital in our position. More importantly, we have a extremely good team in place.
We've been successfully hiring new people and both on the science and technology side, as well as in a sales organization. All in all, of course, we have our mission and vision there. At the end of the day, the patient outcomes matter, so and with our solution, we can help the industry big time. Now I will hand over to Veli-Matti. We're going through the financial side of the-
Stuff
Business. Yeah.
Thank you, Jukka. Always pleasure to present good numbers, which exceed at least my expectations. Our revenue accelerated towards end of last year and ended up at EUR 1.9 million level.
EBITDA was - EUR 8.1 million, and the net loss for the financial year was - EUR 10.6 million. This is all as expected and planned to keep up the speed to stay ahead in the competition. Equity ratio was 85%. Cash was in a healthy level at EUR 24.7 million. Number of employees, as Jukka mentioned, grew to 65 persons on average. We end up having 78 persons as employees at the end of the year.
Besides that, we do have this hired workforce, which end up having the total workforce or team over 100 persons. Earnings per share was - EUR 0.41 per share. Our financials are pretty straightforward. Over 90% of our costs come from our fixed costs. Employee costs, EUR 6.6 million, and other operating costs are EUR 8.7 million. The employee costs will increase this year. We recruited heavily towards end of last year, and therefore, we see some cost increases on that side. On the other hand, on the OpEx side or other operating costs, we think that it's leveling at pretty much these levels. Two-thirds of our operating costs come from software costs like cloud and then hired workforce, software developers mainly.
This hired workforce really gives us flexibility and buffer to cope with the changing needs in our operations. The balance sheet is really light, asset light. We don't have to invest in inventories, machinery premises. The biggest item after cash is R&D development costs. Last year, we activated EUR 5.7 million and then added that with a recognized grants that comes up with total investments in R&D as EUR 5.9 million. The other investments were last year, we invested some the furniture in our new offices we moved last year, plus some computers and laptops. Really light asset balance sheet. Cash flow from operating activities, we promise this to be on a positive figures by the end of 2025.
That comes from the profit of course, but also this changes in working capital. Over here it says that our operations have tied in some capital but in general, it should be a positive figure here, but negative working capital. Our business model is really, it's not tying any capital, it kinda like releases capital. We get license fees up front, then we recognize them throughout the year. Like end of last year, we had EUR 1 million as prepayments in our balance sheet. The year before we had a half a million. When our revenues grow, so will grow our prepayments. That makes it that the changes in working capital would be a positive figure, but it's negative working capital.
Our operations doesn't tie any capital. The board of directors proposal for the AGM is that the loss would be placed in the retained earnings account and then there won't be any dividends paid out. Over to Jukka.
Okay, back to me. That was looking the back mirror from last year, which was very interesting. Of course everybody thinks what happens in the future. I can say immediately that we are not giving a guidance on revenue and profit and that type of things. The reason is really that it's a new opening up market and we are still in a very, very small numbers, relatively small numbers and one or two deals makes a big difference. What is really interesting to follow up that we have a steady flow of new accounts, new customers coming in and how do we win, what type of opportunities we have with those. Of course, we have.
I'll go a little bit later on the midterm and the long-term targets, what we have. What we are focusing on, basically two main things, I think selling and making the product. If we simplify what the company needs to do. We already are told that, okay, when you go to a clinical side, it's really a business process automation case, so that bringing a one AI model, yes, adds nice value. The real benefits are coming, if you can cover the multiple different disease areas on that workflow, and then the customer is getting the best value out of the solution. We are continue with that journey so that we add AI models in our portfolio.
I have to say that we have already + 400 done with our customers and ourselves. The CE-IVD mark, we have five at the moment. We are going to expand the portfolio on that side as well. The other one that I briefly touched already is the kind of a preclinical side. Where I said we have the AstraZenecas and Novartis of the world and multiple other companies who can benefit on a pharma side. Then there's a research side that there's a lot of smaller research groups who can benefit from our kind of a platform.
The platform, what we have to create one is actually the crown jewel of the company because we are using that platform ourselves to build those CE-IVD mark AI models as well. The whole idea was that we enable the pathologists to build their own AI models, and that works in a research setting very well and also in a pharma setting. Pharma companies can also benefit the pre-made AI models. Playing on those two markets and making sure that we have a leading solution offering for both markets. That's core of our existence going forward. Of course, now we are in a phase that I think if you think about the Crossing the Chasm and all those type of things.
We've been playing with the early adopters. I think we are now seeing the early maturity and before going to a kind of a tornado phase. We are getting now the Mayo and such that okay, are the kind of a referenceable best practices for the industry. There has to be some level of a consolidation and simplification happening, it captures the volume. The market is there, market is waiting that offerings are ready to go and start being used in real life as well. Land and expand. That's the idea of course that what now we have sizable, good brand name customers. We've been delivering the first AI models and the idea is now that you can grow.
It's a good strategy for us because there's plenty of things that we can sell to the existing customers and we need to do, of course, the implementation, the cloud implementation, we need to do the security things and moving forward, adding an additional AI models to an existing customer is actually very profitable and a nice way to grow from our perspective. The commercial partnership. We have Epredia. Epredia is the key partner for us. The team has been really started to deliver on a pipeline growing side. We have closed some deals already and there will be new deals coming up on that side.
Epredia is a key player, but then on a technology perspective, we have multiple other companies as well we need to partner with to make sure that we cover the ecosystem correctly there. We strengthen the position in the U.S., but also in Europe. Now we saw the 253% growth in the U.S., 68% of the business coming there. The Europe is coming up as well. We see that the pipeline is growing and my expectation for this year, even though I'm not giving any numbers out, I expect that, okay, Europe start really catching up and performing nicely as well but still going strong with the U.S. and growing that business.
Well, this was a release coming out a couple of weeks ago. It's a topic that we've been talking multiple times already, I think in IPO times, that we were part of the PathLAKE consortium. That's the organization who's dealing with the NHS in U.K. and trying to find a right AI solution for pathology. There was a tender, so that we were competing on, and we won it, and now we are able to sell 25 hospitals. It's actually kind of a, I would call it hunting license type of a deal, so that they've been pre-qualifying the solutions that the hospitals can buy.
It was a very heavy process on that side, so and then we succeeded very well. On a prostate cancer side, so there's three vendors who are now able to sell to these hospitals. We are one of those. Of course, the other, which is even better, is the lung cancer PD-L1 model, so that we are the only vendor who can sell to these hospitals. I think there's two takeaways on this one. First, once again, the solution was qualified and accepted, and we won the deal. The second thing is that, okay, so there's a sizable kind of a selling opportunity with the NHS hospitals and then starting, of course, expanding according to our strategy.
At that hospital group that was covering something like 3.6 million slides, 25 hospitals covering 20 million people area. There's a sizable business available. Once again, putting that in a perspective for the Finnish audience, the whole Finland is a 1.6 million slide. This hospital group is at 3.6. Mayo is at 2.6 million and growing. In a way, there's a lot of big opportunities out there. Quick snapshot on how we are doing with our targets. If I say about 2022 and what are the key takeaways from my perspective. We were growing on a sales side as planned, so nice growth numbers.
On investment side, we were actually doing the things as we promised to do. And the overall umbrella message would be that, okay, we are actually moving forward according to a plan. Going to the plan. We had a plan to make six CE-IVD mark solution before end of 2023. We have five. Europe changed the regulatory bodies half a year ago, so there was some delay, but we have the models ready. I'm confident that before end of the year, we have the six or more on that side. Then five customers on a clinical diagnostic. Yes, we have three. And then another target in a big pharma, nine out of 10.
Yes, I believe that, okay, before end of this year, we have reached all of these first three goals. The last one, over 5,000 users, so we have already achieved that. We are definitely on track achieving all of these short-term business targets that we set up. On a mid-year. This is, of course, getting more challenging. We have the idea of having a good coverage with the different AI models. That's progressing well. Then, as Veli-Matti said, that we have a positive cash flow from operating activities before end of 2025.
All indications, what we have, if everything goes as it seems, we are able to reach this target on that point of a time, then revenue over EUR 100 million. Yes, that's achievable as well, in a connection of a market opening up, us kind of, accessing to our customers and expanding the operations, but, certainly doable. Then reaching 20,000 users. Yes. On track on that. As I said, over 5,000 already. Then focusing on those bigger accounts and winning those. We are well on track on that as well because we have those brand name customers already using us, doing the first proof of concept projects.
Next thing is to do turn them as a real, a big paying customers for us as well. Well, forthcoming events, we have the AGM coming up end of this month and then we have the financial reporting. We, we do the kind of a reports half a year but of course we t ry to keep up regular communication. Let investors to know what's happening in a company but the actual reports every half year. Now there's a time for questions.
Thank you, Jukka. We actually have a question here at the audience in the studio, go ahead.
Okay.
Thanks a lot. Hi, Antti Luiro from Inderes. A question on your growth. I know it's a tough part to say where it's going but obviously you won large customers and I guess fair to say that the state of your or speed of your growth is a function of how fast those are ramped up.
You know, big customers, many locations, many models they need to take into use. You've obviously worked with some clinical customers already. Now, based on the learnings you've gotten so far, how do you see the ramp-up schedule going on a typical customer? I know there's no accurate answer for this. It's a speculation.
As we all know, it's a new market, there's no kind of a standard or how it should happen. For us, Mayo was kind of. It was a heavy uplift to be very transparent, so we put a lot of our resources on supporting that delivery and there was a lot of things coming up like I mentioned, those security things that they required and we did some additional certifications on that. I think those are usable on the next customer, so we don't need to do the same heavy uplifting every time. We did something and we learned and we can reuse that.
Then of course, the cloud is always a topic and cloud security and maybe that's the one thing why the U.S. is ramping up faster because they are fine with the U.S. cloud providers. In Europe we have time to time the discussion: is it secure? Is the U.S. having an access on certain things? Those have been sorted out as well that they are now lot of work done on making sure that the major cloud providers are safe and secure and the hospitals can adapt. It was a little bit delaying the adaptation but back to your question.
When we start delivering like Mayo, now it's done on the research side. Fine, it can expand because more they use, more business we get out of that. That's nice. It's working already. On a clinical side we have the first breast cancer related AI model in a production before end of the quarter. Adding a new models. It's much easier because the basic setup is there already. We start adding a new models. That is a business opportunity for us because of course we charge per case handled. More AI models we deliver, more money we make that and more importantly of course, more value the customer is getting out of it as well.
If I read into that, and try to answer
It was a long answer, sorry.
Yeah.
Yeah.
No, that's really helpful. Maybe the first year could be the ramp-up phase, things are taken into use. When we speak of the next two, three years, that's probably the point when the usage is on the move upwards more clearly.
I think it requires at least a half year to. It's a little bit similar like the ERP business comparing to SAP or Axapta of the world. Once you implement something, you need to of course do a transition from an old way of doing things. You need to train the people, so it's not just the software, but it's the whole process out there doing things. When it's up and running, you don't want to change, so there's a certain level of stickiness on this business model so that if you implement Mayo almost a year as we did, then they are not likely changing that immediately, so that when you get in, you are in.
That's why it's really important to win those deals. Go even with the small one in and start delivering something so and then the customer start growing but we expect that okay, the delivery will be quicker. The second one will definitely with Wake Forest and the City of Hope will be faster than Mayo for sure but okay, so how much faster so I cannot give exact details yet.
Yeah, yeah. I get it. Yeah. Thanks for that added color. Another one on your investment levels. I guess with lots of deals going in the execution is going nicely, when you look at your cash position over the next three years or so, do you find any constraints from your financing situation in terms of how much you're able to invest in growth right now?
Well, as we said in a bulletin as well, that we think that the heavy investments are done. As Veli-Matti said, we increase the employee costs, but it's not likely that going forward we need to increase that the big time. There of course there might be some changes happening, but the level set is done. It's a software business. Even though it's a enterprise software business type of a business, but we can still grow with this organization and scale up, so we don't need to increase people if Mayo buys an additional 10 AI models. We may put one or two supporting people in place but not significant changes.
My point is that we can scale up with the current setup quite nicely. The next part of the question I think is that it seems that the market is opening up big time, so how can we be in a market all over the place and all over the world and capture the opportunity which is very important to do in the next coming years? We have funding enough to grow according to that plan that we showed. If there is an opportunity to accelerate so we can partner better or we can do something on that side, or then some other kind of a financing action. It's not needed.
You really can't tell what will happen in the next coming years, how quickly it's opening up. Do you want, Veli-Matti, to add on?
Yeah, that, I mentioned that on the operating cost side we are leveling up, so to speak, at these levels. Since the main part of our investments is just kinda like activations from that side. You could draw conclusion that the investments are not growing at least as activations or putting the balance sheet. What comes to this, what Jukka mentioned about the deployments, of course, we would need perhaps more hands to do the deployments, but that's not really investments. It's investments for people to execute the orders.
Yeah. It's also helpful. The last one from me. Looking through revenue, I know your revenue is largely usage based overall. If you look at your revenue growth, is that a good proxy of how much the usage of Aiforia is increasing in your customer base? Or is there a significant amount of project revenues in the revenue growth?
Well, it's. Maybe you can jump in with the details, but there's a certain amount, of course, that, okay, when we do the implementation, it's a one-off thing. That, okay, we deliver the project, we do some kind of a bespoke tailoring for the solution. It moves to a pure software type of a recurring revenue after that. How much exactly? It's a mixture of a little bit developing at the same time and delivering at the same time. The majority is definitely coming from a software. Because we sell the kind of a license, and that's an annual fee for the create license.
When it goes to a production in a clinical side, so it's a case-based pricing, so, and then it's pure volume.
I could add on that, I mean, these figures show our revenue, not the invoicing. Invoicing is totally different. We get the licenses, we recognize them on a monthly basis throughout the next year. That also heavily affected on that networking capital thing. We invoiced quite heavily towards end of last year. That's why we did have receivables in our balance sheet, and that tied in some capital.
Yeah, invoicing is bigger than the revenue.
Yeah
Because we don't recognize everything at once.
Good. Thanks. That's helpful. All from me.
Thank you. We have some questions here in the chat as well. First one regarding the existing challenges. Where are your problems?
Well, I think we are still in a small scale as a company, so we need to definitely improve the pipeline. We need to improve our awareness in the global markets. We are well-known in the community, so with the pathologists, so that we are all the events, and we have so many scientific papers done so that the credibility is there. When you start growing, I think there's about 15,000, 16,000 laboratories in the world, so you need to have a feet on the street to be able to do it. And that's the one thing, so that we need to make sure.
Thank you. The next one: Did you book any revenue related to clinical implementations in 2022? If not, when should we expect first revenue recognition from clinical projects?
Could you repeat?
For sure.
Did you book any revenue related to clinical implementations in 2022, so last year?
Yeah.
If not, when should we expect first revenue recognition from clinical projects?
We did. The implementation is a bit different from the recurring revenue. We, those deployments when they are done, so they're booked as revenue and the remaining part as the license and the analysis, analyzed that's booked on a monthly basis or recognized on monthly basis.
Thank you.
We booked both last year also, but it continues this year.
Right. I think we have time for one more question here. How much did you growth revenue pharma customers in 2022?
We don't release this type of numbers. That is all we could tell at now. Then, the answer for that is like, our revenue is so low levels at today. There could be big changes and that like disrupts the whole idea of how our revenue is growing. Perhaps within couple of years we could release this type of figures as well.
Thank you. I think we are done with our time here. I think it's time to wrap it up.
Thanks.
Okay. Thank you very much.