Teneo AI AB (publ)

okay we're going to go ahead and kick off as a reminder this meeting is recorded and you will find it and presentation and the report on Teneo AI investors Teneo AI about investors on our website so with that said i'm very glad to present another quarter of growth for Teneo i also see that we have quite a small crowd today so if you wish we can make this interactive so if you raise your hand as we speak then Then I'll be happy to take questions as we go along as well. Otherwise, we also do have, of course, questions at the end. But just to, since the crowd is quite small. Great quarter. We have some very exciting product development. I'm going to dive into a bit more why they're exciting. We have some concrete customer feedback with real KPIs that are really interesting to see. We've had a few of these over the years, but this is really, really interesting to see how people react to using our technology. And also concretely, I'm going to show why I'm still committed to our $20 million ARR goal at the end of Q1. So without further ado, I'm going to step right into it. As you can see, the presenters are known to almost all of you on the call today. It's me and Fredrik, and we've been at this now for a while, and we are now feeling that we are at the end of a major transition where we're also going to start delivering the new customers that we've all been waiting for for the last year and a half since we started that push. I do want to start however with the market update I think it's quite important to take stock of where the market is and that is the general AI market and then specifically our market the contact center market I'm going to start with the AI market so there's been many many things happening in this space lately and One of the more important ones during Q3 was that the report from MIT, which was later developed upon by Harvard Business Review, you can read it here on this link, which says that 95% of AI projects fail to deliver measurable returns. And we see this. almost all the really big ones in our pipeline that I'm going to talk about later have tried to develop only on top of a model or with a thin layer of veneer on top of a model which I famously call lipstick on a pig which means you just basically have a prompt that I provide I'm the supplier you supply to your customer the customer puts the prompt on top of their rag model and then you hope that everything is going to work it turns out it doesn't and there's many many reasons for this many technical reasons happy to go into those two as some of you who are following this space may have listened to the two hours 20 minute interview with Andrej Karpathy who of course is one of the most proficient users of AI technology today It is not like the AI agents are going to take over tomorrow. If you saw the launch of Atlas this week, or maybe potentially the new Gemini agent builder that was released last night, there's so many things happening at the same time. But are you prepared to let Atlas, that's the new browser from OpenAI, which is brilliant, But are you going to be prepared to let that go in and log in as you and do things on your behalf at this point? I think not. With the hallucinations and the other problems we have, you're not going to do it. Is an enterprise going to do it? No, they're not. So there still needs to be a rigorous enterprise framework, which is where Tenio 8 or Launch really comes into play. This is interesting also because it's of course made by, I believe I made this with Gemini or GPT, can't remember. But it is a market in flux and you see some of the signs and so forth being incorrect. But what this is symbolizing to me is at the bottom you see all these contact center providers and there's loads of them, right? So we focus a lot on Genesis because they have the really large customers in our markets. but there's many, many. I mean, there's a local one in Sweden called Pussel that recently was acquired by a PE. There's all these, there's Odigo in France. There's, I'm in Holland today, there's three local Dutch providers that are contact center providers. And in the US there's a plethora of them. The interesting thing is what the contact center providers are seeing now is that you see the CAI, the conversational AI companies, they're sort of shooting past them. And this is very obvious when it comes to the ones that are going agentic or lipstick on a pig. What they're trying to do is replace the contact center. So they're saying, hey, the phone number comes into my agent. The chat comes into my agent. I'll just automate everything. You don't need a contact center. And of course, that makes these contact center providers very, very nervous. and that's why we saw the nice Cognigy acquisition at a billion euros dollars here recently and we see lots of interest in the market to work with us as OEM but also of course others are coming to us for other reasons than just being an OEM we are committed still to our partnership with Genesis but we're also building out as you'll see in our product update to be able to integrate to any compact center as this market is really opening up and people want to automate that's so far for a market update let's move on quickly to a business update So all KPIs grow year-over-year. The ones that we track and are most important to us is the growth in SaaS, software as a service. The ARR there is growing quite aggressively as year-over-year. You remember we have seasonality, so we're very focused on the quarterly performance year-over-year. And we also saw great revenue growth in the SaaS API calls year-over-year. This is very important to us because that drives to another number, 87% gross margin. which is the most important part for us. We've reduced the cost of delivering the service, but at the same time, the more calls that come through the platform, that's where we make money. API calls, once they're above a certain level at a customer, are just about pure margin. And you'll see why that's important to us and why we're focused on this ARR goal that I'm talking about. We also had a very good net revenue retention. This is still very, very good. You've probably seen other reports, especially in Swedish SaaS, the Motero reports where that is a very, very good number, 121%. Some of the operational highlights obviously spoke about the growth in revenue already and API volumes, but we have a new sales agreement. This is the first of the big ones that we wanted, that we've been talking about in the pipeline. and this one is very interesting. It is essentially a company that is in the HR and staffing world, but what they provide is staffing for things like contact centers as well, and what they're building is software as a service, and they're selling in geographies where people are quite happy to replace the people with the software as a service. So we've done meetings together also with new customers with them, apart from the project of converting existing volumes and existing customers and for obvious reasons as they're in HR we're not going out with more details around what the company is but it is an interesting development it is BPO so business process outsourcing but this was a paper initially from JP Morgan in let's say April time frame It's now been picked up by many others. You talk about instance software as a service, you talk about service as software, meaning you deploy services, but you do it in the form of software. And obviously, with the agentic AI platform Teneo, you can do that in many different ways across phone, chat, email, etc. The other thing is we've renewed quite a few agreements over this period. We have a partner here, Ebo, that you see at the bottom. Quite interesting. They are building a proof point for our customer, NHS, in Qatar in one of their hospitals, which is pretty much populated by UK doctors. It is one of those hospitals where great service is provided to just show that even a VIP hospital really likes to use to pick up the phone and do patient calls, as in, can you come in at 10 o'clock instead of 11 o'clock, or we have a procedure for you tomorrow at 12. And then, of course, the big operational highlight from a product perspective is Tenio 8 that we've been working quite hard on that really addresses some of the things that the market is going to be looking for now with this big contact center being scared of the AI push. Investment on new sales. This is a slide we've used now for the last few quarters. I'm not going to dive into the calculations in one and two, but we do measure the pipeline only with the subscription, not with the API calls. And we measured in probabilities. That's what everybody does, of course. What you see is that it's almost doubling again. So we are really building in a pipeline. We're both progressing the pipeline. but we're also adding more and new into the pipeline. And some important development there, I can just mention FGS. FGS is in Turkey, and we see great interest in our services in Turkey. And there's many reasons for that, but not the least of it, their economy is not doing great, so cost saving is on the table. and very much a very large population that are very tuned to calling because a lot of people don't have smartphones. It's a country that is developing very, very quickly. But of course, as some of you will know, big inflation problems meaning big big cost push there so very interested in the perspectives in Turkey as well and actually one of the big ones in our pipeline is a Turkish but global company they actually are I'm actually in Amsterdam to meet them today but the headquarters is not in Turkey but they are a Turkish company by by origin. Here's a customer. This is Medtronic, one of our customers in the medical field. As you can read in the report, they have Virtue Earl, named after Earl the pacemaker inventor, which is a virtual assistant that goes across many different solutions from diabetes to heart disease and for patient communication. We can skip sort of the above, the ones above. Yes, it's millions. Yes, it's, you know, they save a lot of money. But if you start looking at customer satisfaction, this is plus 8 percentage points. So it's not plus 8%, it's plus 8 percentage points. So if there's CSAT, I'm not sure they didn't disclose what their CSAT was, but let's say it's 50. That means that this is 58%. Let's say it's 90. That means this is 98%. So that is significant. So the customer satisfaction with talking to a bot is higher than a human. We saw that before in transactional MPS, that is, I make a phone call, I get a survey afterwards. This is actual surveys of customers where they ask this specific question. So very, very happy. to see this type of numbers. Obviously, they cut wait time by 37%. That is probably why customers are more happy. I don't think they really like necessarily just talking to the bot instead of a human. I think that the bot solves it, and the fact that it's fast, that is what people really appreciate with these types of solutions. Their abandonment rate is down. Abandonment means somebody calls in and they hang up before they're helped and they improve their service level as measured by the customers by 18 percentage points. So very, very happy to see good KPIs in real surveys done to real customers here as well, which again validates that this is the way to go. So now I'm going to talk about the goal. It seems like a bit of a Superman stretch to reach the goal. So I asked Microsoft Co-pilot to draw this very Superman-esque going here where you see comments and other stuff coming against us as we're reaching for the goal. But I want to quantify why I still stick to the $20 million in Q1 ARR goal. It's not really superhero stuff. So if we start on the right of this slide, you see that existing customers, I calculate in Q1, including the new customer, is going to be $11 million of ARR. So exiting Q1, we're going to have $11 million ARR from existing stock. Now that means we need an additional 0.75 or 0.8 million dollars additional revenue in March. And that means if you break that down into just API calls, there's obviously also subscription revenue. But if you just do API calls, which is not measured in that pipe we just discussed, but this is why we are addressing these customers. We did roughly 80 million API calls, which is roughly 13 million calls per month. I did the 13 to 80 based on the pipeline. Some of our customers have more, some are fewer, depends on what the pipeline is, right? So, let's say we need 80 million API calls, 13 million calls. Now, the current pipeline that we have right now is we have one customer where we are currently in an active pilot. This is an existing customer where we have half the estate, somebody else has half the estate, and we are now in the pilot. We are very secure that we're going to be able to prove that in the pilot that it works. We're also fairly sure that we're going to be able to do the contracting through the existing contract, meaning that we should be able to transact before the end of the quarter. There are 36 million calls per month. the next one is 34 million calls per month again very large volume fairly certain we're going to be able to transact at the year end because they have a budget and we're in negotiation 9 million per month we're already negotiating with procurement we already proved our point on this one So we prove that it works. We have another one of 4 million months and then we have 3.3 million per month. These are the big ones. And as you can see, we could do the nine plus four. We could do the one of the 36 and one of the 34, or we can do eight times eight of the 2 million call per month customers we have. So we have good coverage to see that we can reach this now. The only question here would be timing. And as you may know, I did say I expected Q4 before, but the big thing that has really been pushing these out is that we keep getting these LLM questions that we have to we have to tackle as well. That is getting less and less, especially with Tenio 8 as DLLMs are simply native and you don't have any issue of taking the lipstick on a pig and quickly in a few hours doing exactly the same thing. And so very, very good pipeline, very proud of the team's work here. We have a brilliant, as you may remember, we invested in what we call solution architecture. So we build the pilots ourselves, and that has really progressed the pipeline quite a lot to prove that this is something different than just an LLM with some lipstick on it, which has been the major delay in all these. We were also Genesis Experience in this quarter. That was in Nashville. The containment within this. So this is a captive market we like to call it. So we address the one million plus phone calls per month customers within Genesis. That is our prime market. That's what we do most on marketing. That's where the sales do most of their outbound calling, most of our planning. And then we have others which are more opportunistic inbound things that are not Genesis. But Genesis is our prime, market we're on the genesis app foundry of course and what we see why this market is important is that i have very very low containment it's mostly keypad ivr still and they are less than one percent containment across the board uh they have high abandonment rates etc because it's it's a very large customers very complex environments and difficult to build unless you have to kneel The other AI agents that are on the AB Foundry are at question because they attempt to take revenue away. So they do that where Sierra AI, for example, which is pure LLM play. They try to say, okay, come to Sierra chat first and then we'll just take away the need for phone calls to Genesis. If there's a phone call, we'll manage it on voice. barely but you know we'll try to at least not use genesis because that's the cost so that means that the genesis sales people feel more secure with us we so we have several of these we initiated on site in uh in nashville and then we're very happy to also participate in paris in November, I think it's in three weeks. And that's a similar thing, but obviously for the geography EMEA and Middle East, which is also interesting territories. So now we come to the big public update. I'm going to show a video of some of the benefits of this as well, but I want to highlight some of these benefits. So public APIs, what does that mean? It means that you can address the core functionalities of Teneo from outside from another interface. So let's say that you have a solution built like in your contact center, you have a rudimentary IVR, but you want to make it voice enabled, but you don't want the customer to leave your contact center environment, your development environment. You may be a smaller contact center provider. Now it's possible to OEM and white label us. And the reason we did that is we see all these smaller players, which we cannot address their customer base. That's not profitable for us. But addressing that player that may have 150, 200, 300,000 customers that have five seats, so very small customers, But they have many customers. If we can OEM to them, that's going to be very... And we saw that during the summer immediately after the Cognigy acquisition. We saw lots of these discussions. We're in a few of those. So we've added this in to Tenio 8, where now you could essentially use all the Tenio PowerLift functionality, all the enterprise functionality, but you can use it from your own interface. We also built in the AI agent builder, which makes it very easy to build AI agents so you can have a deterministic business process, but you can use AI agents as parts of it. So let's say that you can use the AI agent to go out and have a long discussion with the customer, capture the keys out of that discussion, put that into Teneo Linguistic Markup Language, TNML, and now you know what the customer is talking about but you have a very pleasant nice conversation or you could just build an agent that goes out and does something like goes into salesforce captures an invoice brings up the invoice pdf reads they have another agent that ocrs it so reads it and then serves that back to the customer Another important thing is Tenio Secrets. That means that sensitive data, not only are we already addressing what's called role-based access. So different people in the project or different business users in the context don't have access to all the data that's already locked down. But now it's even possible to lock away some data. So let's say sensitive passwords. sensitive data, credit card data, et cetera. You can look that down in a separate vault. This is very, very important to customers. And then deferred annotations, which basically means you don't have to rely on very expensive LLM calls. You can basically choose which model to use where and lower both cloud cost and latency based on that. The contact center framework is just a way for these OEMs or others to quickly connect the contact center into Teneo. So we generalized what we did for Teneo AWS Connect, and now pretty much any contact center can leverage our technology. We've seen recent customers that are using Twilio as a contact center, they use Cinch as a contact center, and then this is very important. So prospects and customers. So I'm going to show you a video what's possible with Tenio8. This is our prime marketing video. It is of course on our website, but I'm going to show it to you now. I just asked Freddy to do some video. Just give me a thumbs up that the sound is actually working.

Hi Ms. Jones, this is Tina from Tenayo AI. How can I help you today?

I'm expecting a package to be delivered and it hasn't arrived yet.