Teneo AI AB (publ)

Good morning all. It is 9 o'clock. I'm just going to wait a minute or two as we usually do in this. I will remind you as we're starting that this is being recorded, but you're all on mute and you raise your hand. I have to unmute you before you can say anything.

So. Let me wait until 9.02. Just testing the sound bar. It's okay. I hear you loud and clear. Do you hear me? Yes. Great. I'll just wait another minute and then we'll kick off. Okay, I'm going to start off.

Welcome all to this presentation of our Q4 2025 report. And it's going to be Per and Fredrik presenting here today. And I'd also like to remind you that this is being recorded and will be shared on our website afterwards. And you can find the report on tenioai.com slash investors. So it's been a bit of a tumultuous year. The year went quite well. The pipeline was developing very well and we had good growth throughout the year. And then we had one really adverse event at the end of the year, which affected us financially. And thankfully we've found some ways to, to Perry that, and I'm quite sure that this is just a temporary setback. So I'm going to talk a bit about first the market. then the business, and then Fredrik will go into the numbers. And I want to start with the market because the market has a lot of impact on what's happening. So our pipeline is just as strong. It's actually grown even more since Q4, as well as since Q3 when we last reported. So it is growing quite healthily, but it's also quite difficult to get that final signature on the paper. So enterprise sales is always difficult. It always takes time. and we of course only do enterprise sales, but it's become even more difficult in this environment because of everything happening in the AI industry. So I want to talk a bit about that, but I also want to talk about the fact that we do win the pilots we engage with and with all these new entrants. So we are quite confident that we will close out these businesses in our pipeline as well, although it's been delayed a bit. So let's start with the market update. And to do the market update, I want to start with a very busy slide, which looks like this. This is Teneo and the things that Teneo does. Now we've been developing Teneo for 20 years, but let's say at least the last 10 years, we've been building on the edge cases, the things that you really need to do to get a really good version of user input on the left, all the way over to response. And if you look at this, the purple ones, are the ones that we have built, which are proprietary to us. Many of them are included in our patents. And then you see, and if it's not clear, I'm just going to make it even more clear. This is the portion in our solutions, which is covered by large language models like Gemini, Claude, open AI, et cetera. So the Teneo solution is a much, much bigger solution with much more that's needed to build a good user input to respond. So answering the phone and responding, then just LLMs. At the same time, the market is now full of what I call LLM wrappers, and this has now become quite a ubiquitous term. You hear it from many different actors in the industry that these are the companies that are building an interface on top of an LLM model. And as the LLM models are growing, that interface becomes less and less relevant. But these companies will all need to build all this stuff that you see here, the purple stuff, in order to make a really good solution work in customer support. But many of them are of course promising already that they're doing this, although they just have the yellow stuff and a few of the purple boxes built today. And if we look at those market entrants and how they're being valued in the US, which is also very much part of why the board has decided to do a strategic review, is because these companies are now financing themselves in ways which is just unbelievable if you know what the technology stack is behind these. 11 Labs, of course, is a voice company now moving into our space, Conversational AI. Sierra is a startup company. It was on the board of OpenAI. Brett Taylor started this company, and they are building AI agents where the latest is now that they have an outbound caller that calls and says, would you like to remortgage your house or condo or whatever it is? And you would say yes, and then it connects to the human They call that conversational AI. We call that funny answering machine type technology. Decagon, Parloa, Poly, all of these are entrants into our market now in the last two quarters. And that's been part of the delay in the pipeline. So we've had to fight these in several pilots. We fought all of them except for 11 Labs, essentially. We had a customer join us in September. So the customer resigned in September. They were trying to build with 11 ads, but we haven't really seen them in pilots. But all the others we've been fighting in pilots and won. So we just recently won a reseller partner that's going to work with us. We didn't yet sign the contract with them. We did win the pilot. I mentioned that also in the Q4 report, which you will find on the web. But look at these evaluations. So Eleva Labs 781 million total funding, Sierra 635. This is the cash that they've received. And just about all of it is last year. And most of it is in the Q4. And here's their ARR, which of course they're private. So it's estimates. I seriously doubt, for example, the Sierra number here, that's a number they said they grew from November 30 to December 15, from 100 to 150. If you look at the press releases, and the valuations and then the AR multiples. Now, this is, I'm showing this for two reasons. One, yes, it's why the board has decided on the strategic review. It is why JPMorgan Redeye are interested in working on this together with us because they see, of course, a big gap here in the valuations, but it's also, this is the fear and certainly in doubt that has been hitting us in the market. So, And then I want to talk about some of the patents that we have. So we've updated the patent report that we did back in 2019. And if you compare that, you'll see that the patent valuation is up from US dollars 164 million to 191 million. And then the citations of our patents, and this is really interesting, are up from 19, up from 5.84 to 19.9. We have now a 19.9 average Ford citations per patent per year, whereas the industry average would be 2.05. And this is interesting because it's cited by many, such as Apple, Microsoft, IBM, even JP Morgan, but it's cited by many. But of course, we never really tried to license or use this in any way because it used to be that our competitors were the same as us, sort of entrance in the market, quite cash poor, but not really somebody that will go after anything in the patent space but these patents are very strong and all of them are used by us which means that we are using the technology in live customer implementations which is quite important for the valuation but it's also quite important for the strength of the patent when it comes to other factors and and those other factors are again part of this why the strategic review has been initiated and why why I want to get this market update. So this is an overlap produced. This is produced by Claude, Claude Legal. It is not something that, you know, it doesn't say that we could go after a certain company for a patent infringement, but it's very much an indication of the overlap between our technology, our patents, and what these new entrants are doing in the market. I also added in Microsoft here and Zoom, just to show you that also it's just not the new entrance, but many companies are now working in this space too. On the left side, you have our patents. You have the key functionality, and I'm sorry about the font in this. There's just no way for me to make it big enough because of course it contains a lot more information, this table, than just this, but this is an overview. You have the legend at the bottom, so strong, meaning there's a strong overlap between the technology that these companies are talking about in their technical documentation on the web, and our patents and what we are using at live at customers today. And if you look at Sierra, for example, you see it's a very strong overlap, Decagon, very strong overlap, and PolyAI, strong, Palo, etc. Microsoft, very strong overlap. However, Microsoft sites are patents as well, right? So it's a slightly different case. The interesting thing with Sierra, Decagon, Palo and Poly, is that they don't cite our patents. So they don't acknowledge that we have patents. They also never went out and said that they have any patents themselves in this space. So none of these companies have patents in our space. So Microsoft has patents in our space. So that's slightly different. So this is the background, right? Very high valuations and all of them are trying to build now technology in our space. We have the tech. It's already been proven in the market and we're really going out strong in this market today with that technology. So I want to leave that as a background. We're not saying that we're going after these companies for patent infringements. What I'm saying is we have the tech already and that's why we're winning these pilots. moving into a business update and I want to start with the technology because this is quite strong so we we did talk about this in the Q3 report we launched it during Q4 and it's being rolled out to the customers as well now during Q4 so the there's many many things in this that are interesting for customers but I would really like to call out two of them public APIs That means that now customers can integrate into Teneo into their solution. This is very interesting to somebody like a Zoom, Parloa, Microsoft, or somebody else, because they could take the functionalities we have with the patents we have, integrate that in. And that's also why we have started an OEM go to market motion, primarily in the US today, where we are talking to these types of companies about doing exactly that. And then deferred annotations is also quite important. This means that you can, instead of having lots of... So what, for example, Sierra does to increase their accuracy is they ask many different models about a certain utterance, what that means. So customer says something, what did they really say? If you ask many models, it's going to cost a lot of money. We don't have to do that. We can use our deferred annotation technology and use LLMs once and not again in customer conversations. This is a very, very strong point. This also ties into what we call our hybrid AI technology. So that's the technology where we use machine learning, our own patents, TLML, and we combine that with LLM learning. So the second thing is we launched a lot of industry-specific voice AI solutions. Typically, customers that want to build a complete solution would probably spend about six months. It's quite quick to get up and running. but if you want to build a complete solution for utility for retail etc that took some time so what we decided to do is cut that time so we pre-built solutions where they just have to configure them for their company so in a telco for example it's quite common to ask about accounts so you know my subscriptions you want to have troubleshooting so I have problems with my eSIM I can't get it to work you want to be able to scale across regions etc So with these pre-built flows, it's much faster. And also we pre-built integrations to now build out the flows that you need as a telco, as a retailer, as utility, as a travel company. And that's of course key in our pilots. It's also key in our sales. And then we come to the adverse event, which had a lot of impact and also then has initiated many of the things that have been going on in January and the beginning of February. So we had an original setup with one of our large customers where we could not sell to the customer directly. So we had to have the customer serve via a partner. So this partner didn't have anything in our space. So we did quite a strong contract with them with several non-compete clauses. They were not allowed to sell anything that competed with our solution and neither to this customer, of course, but also not to the broader market. So it was our tech, their delivery. Now they started talking about a new solution that they had in the market already back in May. We made it quite clear to them that, you know, that's great guys, but you still have this contract with us. There's no way that you're taking that to market. They even started now positioning that to our customer. We learned during the summer, we warned them again during Q3 that this is not something that's allowed. And at the same time, it turns out that they also renegotiated the main agreement with the customer to be a service with SLA contract instead of a contract which stipulated to NEO. And what they did, not only did they then replace technology at the customer, they also started targeting other prospects with this new technology, which might even be infringing as in they've copied some of our technology as well. So we could initiate the legal action on this, already right before Christmas, as we were aware of all the things that were happening. The really unhappy thing for the customer here is that the customer now has an inferior solution. And the bad thing for us and our shareholders, you guys, is of course that that revenue is no longer flowing to us, since the customer is no longer using Tenio for these flows, is now using this partner's own solution. You can look at Trustpilot, the partners, the customers. The customer satisfaction scores have really nosedived after this and they're hovering around one. One is the lowest you can have on Trustpilot. So unhappy customer and unhappy customer of customers. But again, impact on us is revenue. We've engaged directly with the customer. We initiated that already directly when we were aware of this in December. and we've had numerous meetings where we now have the specification from them on what they need from us to rebuild the solution for them directly and to supply software to them directly. We are expecting a full recovery of this with a contract that goes direct and thereby getting back to revenue levels but also ARR levels sometime during Q2 of this year. This, of course, had a cash impact that initiated many of the financial things that you also see in the report today. But it also, of course, moves our target of $20 million ARR a bit further out. Strong pipeline. Some of those customers still aren't signing, but we're working on to get them to sign. But then also this direct engagement is going to get some of that, recover that revenue in Q2 that's that's our expectation okay so the financing that I mentioned so what happened then is also because of the cash flow issue and because of the revenue drop we had some our lender has some covenants that we were then potentially breaching so they have been very very very very good and renegotiated these with us waived those covenants so that they don't impact our business anymore because this is now in course of action we will still be renegotiating the loan during Q1 Q2 this year so this means that we the the financial situation is very secure and of course this was done with some key shareholders they want to remain anonymous they basically just gave us a subordinated debt so a loan that can be converted either in the financing round or if there was some kind of M&A event, they would get a 5% 15% interest on this. So it's a loan that is basically, you know, thank you very much shareholders. This is very strong support and this aligns very much that we can now just keep executing on the goals and no longer have a financial financial sort of shadow hanging over over the company. So very happy with that and thank you very much to these key shareholders that took this step to secure our finances and our debt situation. Some of the operational highlights then. It was a year of growth in 2025 with a turbulent end, which initiated, of course, all this communication we've been doing to the market. We launched these industry-specific solutions. You can find them on our website under Solutions. Scroll down to the bottom, you'll find them all. We released Tenio 8, a very important release to us with public APIs and this annotation saves customers money, but also does accuracy. We have the patent valuations and the citations are heavily up. Very, very important. We have this partner that breached the contract, which we anticipate recovering funds from them. So getting damages from them, but also of course, regaining the customer. And we have several pilot wins versus the wrappers, one being a reseller partner that we're going to be signing in the next few days. So we'll announce that separately, but also in three customers and one of them where we're now at the final stages as well of wrapping up the paperwork. Renewal agreements with some very interesting customers, two banks, which is quite important to us. We're seeing that the bank segment is waking up to this type of technology as well. And then the strategic review initiated, of course, on the basis of seeing what's happening in the market right now. I want to just take this last one, which I find very, very important. This is a graph that I've shown before. It is the graph of how the accuracy develops in the live deployment that we do today. This was verified by a company called Sierra. It was done on the banking 77 dataset, which has then been benchmarked by many others. You see some of them in that square at the bottom, Microsoft Dialogflow, it's Google, Cognigy, et cetera. And if you look at the right side, this is what happens in the multi-turn conversation. What's multi-turn? Hi, how can I help you today? Well, I have a problem with my billing. Okay, what billing is it for your broadband? No, it's actually for the eSIM mobile phone okay so if you do six interactions or six turns in a call if you have Amazon Lex as your provider which is part of Amazon Connect or you have Teneo you see at the top here Teneo has a 99 accuracy and understanding and Amazon Lex 89 with six interactions that would put you at 94 understanding with Teneo But 50% understanding index. And this is the real problem in the market. So what do large companies find? We have, of course, one really large customer with a really large volume that is working with dialogue flow. And they find exactly that, that they do not understand what the customer says. The customer gets frustrated and it then has to go off to a human and they can't automate the call. So this is really the key feature still of Teneo. This is the one that differentiates us. This is the one that we push at customers, but also, of course, very much a part of this strategic review. So with that, I'm going to hand it over to Fredrik to go through some of the financials.