Good morning, shareholders, investors. It looks like we have a pretty full house, I think we're actually going to start on time today. I'll probably kick this off as we speak. I want to remind everybody that, first of all, the presentation is available on our investor page. Also, this call is recorded and being transcribed, that you know that. Today we're going to be presenting the Q1 presentation, 2026 from Teneo. The two people presenting is I and Fredrik. Most of you have seen us before, I'm not going to dwell too much on that. I'm going to start with the CEO update. It is a, let's put it this way, not the report we were hoping to publish when it comes to the numbers.
What I really wanna do is give some flavor on what's going on. Is this a car crash? Are we crashing into the railing? Are we going to a low valuation, a low enterprise value, or are we going to a high enterprise value? As you see, I sort of feel that from my perspective, we saved the crash, and we're now heading towards a higher value. That's what I'm gonna be focusing on in my CEO statement. Let's focus on the numbers from my side, and then Fredrik will go through the numbers as we move along.
What I wanna focus on is giving a clear picture of how you should look at the company right now, because I believe it's quite difficult to see that from the outside, as we speak. I wanna give as clear a picture as possible, as also legally possible in some cases. There are some things I'm not allowed to talk about, but the things I am allowed to talk about, I'm gonna talk about as well. There's three paths essentially. I want us to start off by taking stock by looking at the market and what's going on in the AI market overall, and then there's three tracks. There's the ARR target, so our sales of a product. There's the IP path, the patent path. I'll go into that too.
There's the strategic review, which I'm gonna give more flavor to as well in the presentation as we move along. Teneo is, of course, that voice AI model that has been deployed in large enterprises and shown to work in a space where everybody wants to get in as well. The first thing I wanna do is talk a bit about what's happening in the market. There are many new entrants in the market today. We see companies come in and get huge financing checks in that crazy environment in primarily the U.S., or in the U.S., I should say.
It's all about the debate or the question that investors are having and are betting on is, are the LLMs, so the foundational models, the AI models, the Gemini, the Anthropic Claude, the OpenAI ChatGPT, et cetera, will they be able to actually make business decisions and provide accurate answers or not? That's what I wanna shed some light on because there's been a watershed moment in the market in the last few weeks or months that shows, to me at least, that that's not gonna happen and why everybody's gonna need a Teneo-like solution if they wanna provide business processes based on the latest AI technologies. The first one I wanna start off with is a benchmark that was published by Sierra.
Sierra is a company that at this point has received about one and a half billion dollars of financing. They're doing very well in terms of getting attention in the market, and they published a benchmark. This is called the τ-Voice LLM Native Voice Benchmark. What this shows is how the model. This is primarily in the case of Sierra, they use the LLM models only, right? They, they have an application, what I call a wrapper, but maybe that's a bit derogatory. They have an application that then sends the questions to several LLMs, and then the answer comes back, and their application then tries to pool or make the probabilities of which one of these answers is more correct.
When they try this end-to-end benchmark, which they've tested themselves, they came up with 67% end-to-end task correctly done. What that means is, of course, that a third of the tasks are not correctly done. Now, the question is, how much can they increase that, and what's the cost of increasing that to get up to an acceptable level, which is gonna need to be about 90%? We've shown in live production traffic and in testing even higher numbers, but in live production traffic, that we can do end-to-end task completion 91% of the time in live voice deployments. These are not really, they're not apples to apples, but they're looking at the same business process, and I think that's the important thing. Businesses really need to solve a problem.
They don't need the best and coolest technology to solve the problem necessarily. They need to solve the problem. The problem is phone calls are increasing, and they're increasing because humans make more phone calls, but now agents are making phone calls. I don't know if you played around with the latest, for example, Perplexity browser, Comet, but in the Comet, you can even make phone calls now to customer service. If you wanna ask something about your package, you wanna ask about a policy, you wanna do a post a claim, you can do it now with an AI agent as well, which is fairly secure and fairly well-functioning. A bit better than OpenClaw, we're getting there, where humans will have AI agents as well. Phone traffic is going up.
The business problem is we need to automate them because we can't hire enough people to do it. We need AI agents on our side, and that's the gap that we provide. We provide that accuracy, and that's why customers now, after testing all these other solutions, are coming back to us. That's also why we believe that our solution is important to these new players, the very well-financed players, and our patents are important to them as well. That's the first example. The second example is a thing that's lacking, except for the accuracy that's lacking in these solutions today, is the auditability. The ability to see what has happened, the security, and then the certification.
Things like PCI, ISO, SOC, these are standards that you need to follow and in some case certify against to ensure that your solution does the things that it says that it's doing according to the process that you define. PCI being very important, for example. If you have a discussion and a credit card data goes from you to OpenAI or from you to Claude, then you are, according to PCI, the one that's liable and the one that would have to pay damages for that. One of the more competent organizations, I would say, in this in this environment right now is McKinsey. McKinsey, of course, they bought the what's called QuantumBlack company, which are AI experts, really solid, real heavy-duty consultants.
They've been building internally an assistant called Lilli, which is also what they're selling to customers as an internal assistant. This is the assistant where you would be able to access all your information in your company and talk about it and build things and build presentations, build, you know, documents, etc., for your customers. The interesting thing with that is that it contains everything, including confidential things like M&A discussions with customers or anything under NDA essentially goes into this database, and then it's accessible by Lilli internally. It's now been discovered, again, I think this is now two weeks ago, maybe three weeks ago, company called CodeWall, they used an autonomous agent to just see, can I get access to all this information in McKinsey?
It was very easy to do it through a very simple thing. It's called an SQL injection, something we've been doing for the last, what? Almost 20 years, or maybe even 30 years. It came out in the early Internet ages. Because all this is built with AI, that is a black box, right? All the LLMs are black box. Nobody really noticed this within McKinsey, that it was easy to get to the endpoint through an SQL injection. It's getting a bit technical, but I think you'll understand that really what I'm saying here is McKinsey exposed all data internally, all client data, all their internal data, because they put it all in one solution, that database then was accessible through an LLM-based AI agent. This has been going on since late 2024, when they deployed Lilli.
Nobody really knows how much has been leaked. Now there's lawsuits and all sorts of things happening and looking at who's gonna go to jail for insider trading and etc., etc. Very difficult legal space to work in. Of course, they shut down Lilli, which is a solution that they've also sold to companies as an internal bot. This is really difficult material, right? As soon as you let the LLM be a large portion of your runtime, this becomes very difficult. We use LLMs a lot, but we do not use them for the things where compliance and auditability and security is important. That's a real key, and that's what we call hybrid AI.
Then the second proof that things are happening in this space where the models are not really gonna fix this stuff with the 67% accuracy and trend, is that OpenAI and Anthropic, they've just both committed SEK 10 billion in the case of OpenAI and SEK one and a half billion in the case of Anthropic. Anthropic is Claude and OpenAI is ChatGPT. They've committed that to build service organizations that can build all the stuff that Teneo has for organizations, so organizations can deploy the LLM models. This is, this is a very important statement by them. They're essentially saying, "No, the models are not gonna catch up. The models are not gonna go from being probabilistic. You need to build the scaffolding or you need to build the house.
You need to build a whole building, or you need to build a car, and we have the injection. That's it. We're not gonna help you with forward-deployed engineers. They're gonna be deploying engineers to build this, the next generation of Lilli internally. Today, a lot of that, of course, is focused on coding. That's also what's happened in the last few weeks, that the bots were initially focused on customer support, that's our space, and now a lot focused on coding. That's why we also see customers now finally coming back saying, "Ooh, actually didn't work." It didn't work very well, this whole customer care space. Because you need the accuracy. You cannot tell a customer that they owe you $126,000 when they owe you $294,000, for example.
Marc Nachmann at Goldman said, "There's a big shortage of people who know how to apply these tools into businesses and then transform them." I think that's an understatement because it's very difficult to do that. However, if you do it with Teneo, it becomes a lot easier and you would provide the value quite quickly. Next slide is quite busy, so I warn you up ahead, but I do wanna show this. It is the architecture that Teneo works in to show you where we sit in relation to all of this. All that I've talked about here is what we call layer 3, and that's the orchestration of the intelligence in the company. That's where Teneo sits. You have an end user on the left side of this slide.
The end user comes in through whatever tool that might be, Messenger, WhatsApp or the phone or a phone call on FaceTime or whatever it might be. It comes in through what we call the omni-channel layer. This is where we cooperate with companies like Vonage, AVOXI, et cetera, to take that and bring it into our platform. As we do that, we have an STT that we also take mostly from Azure, from Microsoft, and now we have a text string. That text string has incorrections in it, so we take that into our engine, and we correct those imperfections, and then we put it into what we call TLML. This is where the hybrid AI part comes in. That's the Teneo Linguistic Modeling Language. This is where our patents start to play, right?
Understanding what the customer really said is important. This is where we bring it up to 99% accuracy instead of 89%, 90% that you would get from just a transcription. We then run the business logic and look at the automations that can be deployed on this. We use the tools that we have access to, the knowledge base that we have access to. Important here is whoever that speaker is, that customer, they get access only to the authenticated knowledge they should have access to. That's managed by Teneo. Then, of course, we deploy the Accuracy Booster on all this to ensure that we keep the accuracy through the whole process.
If Teneo cannot answer, so in a few % cases, Teneo will not be able to answer and automate, then we go to hand over to an AI agent, and then it goes over to layer 4 to the contact center, and from there into the contact center agent in the back end. However, if we can, we also gonna use the LLM models, and we're gonna generate the response so that the response sounds very natural and flowing. The conversation feels natural. We're gonna use that, but we're gonna inject it again with the Accuracy Booster to ensure when the conversation goes back to the customer, it's a 100% correct response. Then we integrate with all the back-end systems, such as SAP, Salesforce, ServiceNow, HubSpot, Power BI, et cetera. All this leads to quick time to value.
You would have to build everything in layer 3 yourself today if you wanna deploy something else. All these new entrants that are infringing on our patents would need to build it too. That's important, I think, to understand for the M&A process, that today, what we provide is not competing at all with layer 4 and layer 5. We are the orchestration layer, which is what everybody now starts to realize they will need. That's the background on the market. I'm sorry if it got a bit technical. Very happy to answer questions on this, also on email or whatever, later if you wish to understand this more. This is quite important as you look into our 3 lenses. I'm saying we're still keeping our ARR target of $200 million with the current cost base.
Why am I saying that? It's because our pipeline is still strong. It's actually stronger than last time we reported it, which was in September. We actually forgot to report it in December because of this partner conflict that had started in December and has still been playing out in through Q1. We still retain that. We have the other option, which is patent licensing. Patents are, and the technology is also part of the arsenal, the weapons for the M&A discussions. I call it M&A here now. I will dive a bit deeper into that. There are deep discussions with a number of top buyers around what we do and what Teneo would provide to them, and that is, of course, from the strategic review.
Those are the three paths I wanna talk about, and I'm gonna start with the first one. I've shown this slide, but like I said, we forgot to show it in December, but we did in September. From the Q3 report. The investment in new sales is essentially a sale that we do. We only count the subscription revenue, the revenue for the platform, which manages the solution over time. You build a solution, deploy it, but then you manage over time in the platform. The EUR 6.9 million total euro total pipeline, which is a weighted pipeline, is of course a lot larger than it was in September. What that implies is that we're not losing, we are adding more into the pipeline.
What's happening in the pipeline is customers are testing the LLM solutions. They're testing either one of the LLM wrappers, like a Sierra, like a Genesys AI, which is also built on an LLM wrapper solution. They're testing Parloa or anyone, any of the others that got so much financing that they have all the marketing, but they're failing with it. I'll come back a bit more to that, why that is and why these are keep in our pipeline, but it's the longest sales cycle I've ever seen. I've been in enterprise sales since, well, since last century, and I've never seen anything like this, where customers have such difficulty making decisions. Six late-stage deals. These are deals where we're in pilot POCs.
We also have one deal that just closed, which is not reported in, not taken out of the pipeline here yet. That is because it's close to the partner. The partner has signed up a Teneo solution for our, for the largest, one of the largest, I'm not gonna say which one it is, so one of the largest telcos in Turkey. That one, that one only in 18 months, according to the implementation plans that the partner has, will be generating $1 million per month. We are also closing these out, but it is really, really, really long to get that closed out. That's one of the paths to unlock the value to make sure the car goes right, doesn't crash, and doesn't go to the left. The second one is the patents.
I'm gonna start with showing a patent process here. The way these things happen is you start with noticing that people are infringing. All the LLM wrappers, all the ones you see that are talking about doing customer care with AI, all of those companies today are essentially infringing, and they have no patents in this space. You do the DD. We've done that. That's where you confirm that the patents are actually being infringed on. We call that a claims chart. You take the technical documentation of the company, and then you compare that to the claims, and that's done by a patent attorney. That we have already done. The patent attorney is not the litigation attorney, so it's not the one that then starts the sort of negotiation. That's a second line attorney.
We just retained the best attorney in the United States for this. This is, of course, the United States things that we're going after because that's where the money is. That's where these competitors are that do not have the tech nor the patent. This is a dual approach, right? We're looking at approaching them to talk about licensing and using our tech. At the same time, of course, it's something that goes into the strategic review. If you don't have the tech, it's easier to acquire the tech, and it's easier to also have the patents as defensibility going forward. You send the letter. We started sending the first letter last Friday. As you send that letter out to a company, that letter then gets treated in there, hopefully you never get to filing a complaint, step 4.
Instead, you go to the other one, you start negotiating on the settlement, which would be a license fee or some other kind of business construct. The first letter we're sending is very, very nice, talking about cooperation, et cetera. It is definitely built on a real, a real claims chart by a patent attorney. Again, the new attorneys are also the ones that are gonna be now taking over the partner conflict. The attorneys we had during Q1 got very caught up in the tariff debacle that happened in the U.S. and have been very short-shifted on time for us. The new attorneys, which again are number one in the U.S., it is definitely a top-tier law firm.
The new attorneys will take on the partner conflict, which I'm not allowed to talk about more than that, and the patent litigation. Also, which is important, this is what we're discussing to go on a contingency basis. That means that we don't pay until there's a settlement, and then we pay part of the settlement to the attorney. Important to understand those points. There are infringements. We started the process. The potential is for M&A or for licensing. We're using this as an argument for both, right? The patents we're talking about that are the prime suspects or the one that has the sort of heaviest infringements are these patents. I'm not gonna read the patent numbers. The first one is natural language intelligence. Customers speak, of course, natural language. Sorry.
They can speak it in any language. In our patents, we already have, say, no rigid menus, so no IVRs, right? No scripted paths, Teneo instantly will select the right resolution path. That is exactly what everybody wants to provide to the market. We have the multi-agent orchestration. This is where agents can talk to each other without losing the context. Because agents cannot do that natively, you need a layer above that. That's called the orchestration layer. That's another one of our patents. Natural language analytics.
Once you have all these customer dialogues going, where you're calling into Microsoft to solving a support issue, once that's happened, you want to be able to mine that knowledge, and that's when it goes into our TLML, and it's very easily accessible to see why are customers calling, what if we change that on our website, what happens then, et cetera, et cetera. Those three patents are the most important. Here's the chart of how strong the case is for four different companies, and I'm not mentioning these again to, since the attorney has told us not to. The no-name is the only one on here that we might want to do something legally with as well, which is not an LLM wrapper. I'm not gonna say more than that. The key functionalities, of course, are the three patents from left to right.
Those are from top to bottom here. It is very strong overlap in all these LLM wrappers, all these startups that have hundreds of millions of dollars now to start building the technology they need. There is only moderate on the network side for one of them, but that is not the one we compete with on a daily basis, but that is a different conflict, without saying more than that. The next thing I want to talk about is the strategic review. Also important to note, J.P. Morgan, of course, is the number one tech M&A advisor in the world. There is some seriously heavy-duty people working on that together with us.
Of course, Redeye is the number 1 in the Nordics and has done several transactions in the Nordics and provides that Nordic spice that is needed and all the knowledge that is needed from a Nordic perspective in this, in this strategic re-review process. The strategic review process, I wanna share a few things around this because it has to do with financing, it has to do with the market interest as well. Let's start off with it is on track. It is taking some time, but it wasn't expected really to be much faster than that. There's high interest in the Teneo AI technology. This is a space that's hot. A lot of companies are looking at what to do now, especially since this has happened in February, March that I mentioned in the market.
It's the expansion of the service team for OpenAI and Claude. It's the McKinsey, Lilli, big, big, data breach. Then, of course, you have the 67%, 68%, 69% that we've seen on the τ-Voice index from several of the LLM wrappers. The primary goal is exit here. Today, the board has been shifting. The primary goal has gone sort of from finding a distribution channel as well to really doing an exit based on that market interest. We believe that that is the way to maximize shareholder value today. To go this alone, at the bottom right here, the standalone path is gonna require more, more than just interim financing. It's gonna require a lot more money coming in from shareholders.
What's happening in the market is that there's so much revenue spend, so much spend on the other, the LLM wrappers, that in order to get ahead of them, we would need a larger sale cadre, and we would need a larger, more marketing. We could get to the $20 million ARR with the current cost base, but if you really wanna utilize the momentum in the market, you would need to bring in more money. That's why the primary goal is exit. Then, there will be an AGM announced. We have initiated the debt financing. Sorry, I need to mention that first. The current debt, of course, matures in end 2026. As you know, we have a covenant holiday expiring during the summer, which is contingent on and working towards that exit goal, right?
That's all connected. There is an AGM where we will also announce a financing plan in order to conclude the strategic review. That financing can come from, at this point, from shareholders or from one of those attorney-led discussions. I think, and it's very important that I phrase myself in the same way, but it can come from different things. That revenue shortfall, there is a discussion on the other side of that revenue shortfall, which of course comes from one partner that we have discussed before, but which I'm not mentioning more than that. I'm going to move over to a summary before I leave it over to financials and talk about the numbers, which again, is a depressing thing. That's why I wanted to give you this sort of framing before that.
I think it's very important that we understand where we are so that from a shareholder perspective, you can make your decisions fully informed. Near term, definitely headwinds, right. I mean, I would never have expected to be able to show anything else than growth. And of course, it's the absolute opposite, but it's still a very strong foundation. The revenue is from this loss of the key customer, which is a partner dispute. If we look at the underlying business, the margins remain strong, so it's still the same thing. If we add revenue, it contributes a lot to our cash flow. The underlying business is sound. We do, of course, need more revenue, and that's what we're working on. The SEK 20 million is still our target.
It's not going to be achieved within this fiscal year anymore. It's going to be something that we need to look at to have in Q1 or Q2. It is again dependent on that attorney-led discussions, putting it correctly again. The pipeline is definitely on track. It's also, this is important, right? Enterprises are now coming back to us after internal builds or builds with LLM wrappers, so these new entrants. Maybe we call them LLM-based companies going forward. As they fail to deliver, we're getting back. One of those is the project FTS launched. FTS is our partner in one of our partners, and they launched a first Teneo project now with a major telco in Turkey. That project alone is projected to, in 18 months, be generating $1 million of revenue.
That is with the growth of a very large volume, and it's with the growth of the project there. That is the first one that sort of come back from trying with something else. This is a discussion that's been ongoing for probably 9 to 10 months, and have seen that they couldn't really do what they wanted to do with the LLM-based approach. There, Sierra AI, I'm gonna mention the name here, one of their flagship customers invited Teneo, 2 weeks ago and called us and said, "Guys, you need to come back and have a dialogue with us because it's failing on the voice AI side. We need voice AI to work. We budgeted now having fewer people. Well, actually not.
We budgeted having a larger volume with the same amount of people, and it's not working the way it should. Now we're in dialogue with them to create the pilot to show that we can actually solve the problem that Sierra did not. It's all about the accuracy. It's all about making sure that the agent gives the right answer but also understands the question. Very importantly as well, I'm announcing this in the quarterly report, only as an investor news. It will be announced to the wider market, so to customers after EXL's investor days, which are later this week. EXL is a BPO, BPS company based on AI.
It's a U.S., it's a U.S. company that became partner of the year for Genesys last year as well, so they were heavily into contact centers and providing technology into contact centers. They selected Teneo for voice AI delivery after testing several of the new solutions, what I call the LLM wrappers again, but also testing, of course, the Genesys native solution, which is also an LLM-based approach. The hybrid AI won because they needed the effects of that. I'm very proud of that. It has taken us a long time there as well to be benchmarked against all these different solutions. We came out on top, signed the partnership agreement, and started training them last week.
Again, there will be a market announcement which is more, that shows a bit more of that, for customers and for the market in general. I wanted to share it here since we did sign it, signed it about a week ago and started training them a week ago. The IP and legal, it is a top-tier counsel, this, right? This is number 1, New York-based. They will both take over the partner discussion, so this partner conflict that we have, which is impacting the revenue. That's gonna put more speed onto that. Like I said, the attorneys we had were very good but didn't really have the time at this point to support us.
The scope is also expanding to patent licensing, where we're planning a contingency engagement. That would be where we don't pay them upfront, but we pay them with the proceeds from licensing and potential damages that might come out of that conflict. The IP and the patents are also important for the M&A discussions. They are intensifying. It is something that I end up spending quite a lot of time on. It's of course discussions of where our technology fits in this new space, in that orchestration layer, in layer 3 that we talked about. With that, I'm sorry if it got too technical, but just ask me questions afterwards. With that, I'm gonna hand it over to Fredrik to go for the financials.
Thank you, Per. Let's move to the first slide there. You will all seen the numbers. Let me first, I mean, just explain a bit, you know, what has happened. The deviation in sales and SaaS ARR in the first quarter 2026 is, to a large extent, linked to the decline in sales from a former reselling partner. Per has mentioned that quite a bit earlier in the presentation as well. As we also mentioned, counsel discussions are underway due to this and where we also will be pursuing patent licensing deals with organizations that are actually infringing on our patents. On the positive note, the company's pipeline for new customer is continuing to develop favorably.
We have a record high pipeline, as you could see on the previous slide presented by Per. Teneo is progressing the sales pipeline and continues to have a strong momentum with new customers. Intention is obviously to replace lost revenues here in Q1 with new customers coming in. What we also see is obviously, as for many Swedish companies selling abroad, currency has had a significant negative impact on our ARR and revenues during the first quarter 2026. Year-over-year, we have a 9% negative impact on net sales. Net sales in constant currency amounted to SEK 12.3 million versus SEK 25.3 million Q1 2025, so a 51 decline in sales.
SaaS ARR in constant currency amounted to SEK 27.5 million versus SEK 75 million last year, a 63% decline. Significant drops as you can tell as well. Even with this decline, we also record a stable gross margin, which also show that we have very good scalability in our operating model underlying. We reported 80% gross margin despite the volume decline in the quarter. On cost side, I will come back a bit on that. Going forward, we will continue to invest in sales and marketing, and also, linked to that, we will also have some investments related to the legal pursuits that Per has mentioned quite extensively here earlier in the call.
We still keep cost control, so we have control of all the costs and follow and track this continuously, obviously. Adjusted EBITDA in Q1 2026 amounted to minus SEK 12 million, significant decline due to mainly the volume drop in the quarter. Positively also in Q1, we secured SEK 25 million in subordinated debt from key shareholders, and we are very grateful for that support from these key shareholders. I think we can move to the next slide, Per. API call volumes on SaaS is obviously a key indicator of how our business is doing. Unfortunately, we are not growing in the quarter, as can be seen from the graph on the slide.
This KPI is obviously an indicator of how our customer application and the usage of them are growing. The more applications, solutions, covered regions, et cetera, the higher API call volumes. We recognize a decline quarter-over-quarter as well as year-over-year. As mentioned earlier, new volumes are on the way from our pipeline as we close new customers and new projects with existing customers. On that also, once again, mention the deal that we are pursuing now in Turkey as one positive example that we can highlight as well. As Per also mentioned, the pipeline is very strong, and we have a number of long-developed deals that we are continuing to work on also finalizing.
We see a lot of faith in the pipeline going forward. I think we can go to the next slide, Per. As already said, currency had a negative impact in our ARR in the first quarter. In that regard, I think it's very important also given the decline we have on volumes to highlight the currency-adjusted growth rates. SaaS ARR in constant currency amounted to SEK 28 million, as mentioned earlier, versus SEK 75 million last year, a 63% decline. Similar negative growth rates on total ARR as well. I think we can move to the next slide, Per. Gross margin, as we have presented as a key metric earlier, it's stable despite the volume decline, 8% in Q1 2026.
Obviously higher SaaS API call volumes is important for us to continue to expand the gross margins to levels we have recorded previously and also where we can further improve as well with higher volumes. Next slide, Per. On the OpEx side, just want to acknowledge that in Q1 2025, we guided for somewhat increased cost as we carefully continue to invest in sales and marketing activities. We have done that since then, and we'll continue on that path as well. Despite this, we have also managed to keep cost control, so we actually have a lower OpEx run rate Q1 2026 versus 2025. Essentially, we are keeping cost control.
As also earlier mentioned, we will continue to carefully invest in sales and marketing since we also have a very strong pipeline and interest in our underlying technology from partners and end customers as well. Still obviously keep cost under tight control. As Per mentioned, we have retained a new top-tier legal counsel within contingency engagements. Essentially where we are doing a kind of sort of a revenue share agreement based on outcome. This will cost some money, but less money than if we would have pursued this on a standalone basis, obviously. With that, I think we can go to the next slide, Per. Cash position. We started the quarter with a strong cash position.
We had a cash position including collected receivables, so an adjusted cash position of SEK 29 million at the beginning of the quarter. Also as mentioned in the earnings release, and also brought up here by Per, the company plans to secure financing for ensuring there is sufficient time to conclude on the strategic review process. Part of that is obviously where we have different alternatives. One is obviously that we can go the counsel-led pursuit where we would potentially end up in a settlement situation where we could finance the company and the strategic review process that way.
The other alternative would also be, which I think Per did not mention, and that is obviously that we could lend money from our external parties, or key shareholders to finance this pursuit of the strategic review process. Thirdly, obviously, there is an alternative to also go to all the shareholders to ask for money. This has not been concluded yet on the way forward, in how we will ensure that the financing will be secured for this process to be reviewed. We will come back on that later on here in May. I think with that, I'll leave over to you, Per.
Great. Thank you, Fredrik. I wouldn't really wanna be an analyst and try to model this right now, but I will again try to summarize and give you a good, as good as possible picture of where we stand right now. Obviously the revenue impact is big from this. There's so many things you could say about this, but I'm not allowed to, because of the essentially our attorneys have told us not to. We still have the capacity to deliver with the current cost base $20 million ARR, which of course would make us very cash flow positive. That is one path that the company can still take.
If we would also like to expand that even more and take momentum, of course we would need a different financing regimen, since we are competing with companies who are very well-financed from a sales and marketing perspective. We've seen a lot of competitive validation where the LLM wrappers are failing. Three of the more important ones here is the FGS project that they won after the major telco, working with another one of those wrappers. We saw one of Sierra's flagship customers invited Teneo to come in and fix the voice AI, or rather replace the solution with something that works. EXL selected us, a very important partner for us since they are a partner in the contact center space.
What's happening in contact center, which maybe I didn't talk too much about when I showed that architecture, is that since now only the calls need to go to human go to the contact center, since we're front-running in companies like Genesys and Five9, et cetera, are not too interested in working with us. They view us as competitors. EXL is because they have those customers which then would reduce the payment that they have to pay to one of the contact centers. And then of course, the M&A discussions are intensifying and that is something that is definitely a strong interest for the technology itself. And that's probably understandable with the rest of what I spoke about here. With that, I'm gonna leave it over to Q&A.
As usual, there is a double mute, so you raise your hand, I unmute you, and then you unmute yourself after that if you wish to ask a question. There's so many people on here, so I'll see if I can find
Per, yes, I think we have a question in the chat here from Morten, and I can respond on that one. Essentially the question is, in earlier reports we have written that we.
Yeah, I see.
Expect to win back a lost customer.
I would say that this is very much our expectation.
We, Fredrik, the reason it's not in there is we're not commenting that, so we've been advised to not talk about that. Morten, I'm sorry to say I've been advised to not. It's the U.S. is just complicated. There are so many things you can say, but we've been advised to not talk about that.
Yeah. On that note, I think it's also worth mentioning that Per mentioned the FGS partner or the partner that we have signed in Turkey. The potential in that deal is more or less $1 million monthly. I think it was in our Q3 report in 2025 where we also highlighted, you know, how large the customer deals that we actually are working on and the impact it can have on our revenues. I think there it's also important to note that we are still continuing a lot on working on those large deals as part of also being in our pipeline that Per described.
The potential is very much, you know, that we can quite quickly, once we sign a deal, we can actually have a revenue upside, coming very quickly in our P&L. I think that's the key here.
Okay, let's see. If you raise your hand, I will unmute you and then you unmute yourself after that. Are you saying that Fredrik and I made a very clear case here, so there's no questions? Okay. If there's no further questions, then I would just like to. Oh, here we have a question. Let me open Thomas here. I need to find you on the list first. Thomas, Thomas. If anybody finds Thomas before me, yes, Fredrik, for example, you just hit unmute on him. There he is. Yeah. Now I've unmuted you, Thomas. You need to unmute yourself, and then you can ask.
Yes. Good morning. Can you hear me?
Yes.
Okay. Good morning. Thanks for a great presentation, Per and Fredrik. Just a quick question on the Turkish client so that I understand. The implementation period, is that 18 months? Is that correct?
Not necessarily. It's a scaling period, I would say, is 18 months.
Okay. You're talking about $1 million per month in income. Is that correct?
That's correct. That's when it's fully built out. Given that, they run according to the plan, and it's somewhat easier than others we've seen, there's less unions and other things in that geography.
Okay. Let's assume, if now it takes, let's say, 12 to 18 months to implement the $1 million per month, is that going to bring you to a cash flow positive situation?
I, we need to come back to the $20 million ARR. It's dependent on this. I don't wanna state a forecasted goal here. I don't wanna state a goal here, because it's very heavily dependent on this attorney discussions, right?
Okay. Thank you very much.
Yeah, I mean, $1 million per month is by itself cash flow, would be cash flow positive. Right?
Also to add there, Thomas, I mean, we stated clearly, right, that our OPEX run rate cost, I mean, excluding then variables and also cost of sales is SEK 9 million per month. You can obviously do your math yourself more or less that if fully built out, yeah, SEK 9 million is very close to $1 million.
Okay. Thank you very much.
There is life in the pipeline here for sure. But, like I said, it has been going quite slowly. However, Chandan, can you provide some color on the ongoing JPMorgan review? The project with JPMorgan and Redeye is progressing. We are having, we're now at the stage, which I would call technical due diligence stage. We're presenting very deep technical dialogues with a few of these companies. And that's sort of expected that that's a part of the process. And when it comes to the financials, we are publicly traded, so it's not that difficult to sort of look at our. We don't have any, we don't have any lik i garderoben, as we say in Swedish.
I'm not sure what the English expression would be, there's no real surprises in our financials. We've always paid our taxes and everything. It's really about the tech, and that's the stage we're at. We have a question from Alexander. Let me find Alexander in the list. Teams actually puts the people that raise their hand at the top. Alexander, I'm allowing your mic, and now you need to unmute yourself, and then you can speak.
I believe it was March 18 where Microsoft was publicly one of the key customer reference cases. To what extent can you comment on this question? Is the business relation with Microsoft as a core end user or key customer intact, or is that also impacted or affected by this litigation-related issues?
That's almost like the Elon Musk asking the attorney, "Have you stopped beating your wife?" like he did yesterday in the court case on OpenAI. I believe I cannot comment that. What I can say is that Microsoft is one of our best references when it comes to talking about how the technology provides great value. We are also selling still a majority of our late stage. The 6 that we are in proposal stages are actually through Microsoft. Microsoft selling that's probably what I can say. I am, yeah. I'm sorry that I It's difficult to understand the U.S. system, yeah, we're not allowed to talk more than that. Any other questions?
I tried to provide as clear a picture as possible. I hope that we conveyed it as well as possible. Due to the ongoing process there, we really cannot comment more, which is kind of a shame because there's so many things I'd like to say about that, but that's where we stand. Thank you all. Thank you for your continued support. A bit of a rocky ride here the last few months, but I do believe we're getting closer and closer to the market understanding that you need scaffolding. It's not gonna be, it's not gonna be the steel and concrete that wins. It's gonna be the ones that can actually implement this technology in the enterprise.
Certainly, OpenAI and Anthropic are showing that with saying that they're gonna put all these services people in. Very happy that we that you are still supporting us and listening in. And we're gonna work our best to make one of those three paths provide the highest possible value for shareholders. Thank you all, and have a great day.