Okay, I think, we're gonna go ahead and start, 'cause I'm gonna just give a thumbs up. Like, Niklas, I see you. Can you give a thumbs up if you hear me well? Excellent. Thank you. Great, you can find the presentation as PDF on our investors page, which, of course, is artificialsolutions.com/investors. Then I'm also gonna be sharing it here in the Teams channel. There's also a recording of this Teams, in case somebody would like to see it later, we can organize that, too. Yet another quarter of really strong growth.
We are in the middle of a bit of a tornado when it comes to artificial intelligence, and I can assure you that we are in the middle of it, I, we know exactly what's going on, and we're working with all the right parameters in the new space that's has popped up as well. The big question that I have for this year is new customers, and we are building a very solid pipeline of new customers as we speak. It's really looking promising with our new positioning, which is OpenQuestion as a lead into customers. I'm gonna go in a bit more into that, but just wanted to share that the team has presented yet another great quarter, great management of our existing customers, and really, really good growth. Presenters today: me, Per Ottosson.
I've been in this space for a long time. I am obviously quite old then, I had to have big experience, both from AI, but also from the space of conversational AI, which is the space we're in. Fredrik, our CFO, lots of experience from private equity and SaaS companies, and also from Pareto. A great experienced management team presenting to you guys today. Any questions are welcome. This is Artificial Solutions for those of you getting a bit of a crash course for those of you who are new to this. I'm gonna start up at the top, the top left. It's a very busy slide, of course. We're founded in 2001, if we look at valuations of companies that are founded this year in this space, they're exorbitant.
It's actually good to have experience in this space, so I'm a bit surprised about that. The space of GPT is no different than, the space of conversational or natural language applications that we've seen before. It's a different way of processing. It's a different injection. You used to have, maybe, more of a, old-style, carburetor, eller forgassere, and now we have injection, but it's still the same thing to build a car, pretty much on, on the outside. So very interesting for us that we have such long experience. We have patents, very strong technology platform in this space. Our solution that we take to market, OpenQuestion, optimizes routing. We call it intelligent routing for the contact center to better serve our customers and, and reduce the cost of serving the customers.
It's a very, a very strong offering in that space with the biggest references in the world. Some of these, like Telefónica, doing millions of calls per month, where the system picks up the phone. Teneo is unique in that it's built from the beginning to be language-agnostic, so it runs 86 languages and dialects. We have in operation roughly 18 languages today. That's expanding as we speak, and some of these customers, as they're expanding, they're also adding new languages, which of course, which is a revenue driver. Teneo, the base platform, allows humans to communicate with applications through voice, speech, and text. What we sell is the OpenQuestion on top of Teneo, because that's our beachhead into the customer.
Once the customer has taken that into their heart, they can start doing the other things as well and build more use cases. Teneo can leverage either GPT as an engine or CLU or Google Dialogflow, Amazon Lex, or any of the other AI engines out there in supporting our customers. It's very easy to integrate whatever AI technologies are out there. Before we take the bottom left, I'll just move to the bottom right, sorry, top right. Great customers, great references. I'll go into more, few more quotes on this as well, but, very stable customer base, growing quite, quite well, as we will see as well, their usage of the platform. Again, that's a prime revenue driver and also, of course, a driver of gross margin.
At the partner side, again, the top right, you see three partners at the top. These are ecosystems that we're now a part of. That was an important goal that we had this year, in this quarter, was to become a part of the Google and AWS cloud offerings as well, and not just the Azure. What that means is that if a customer is developing a contact center based on AWS or on Google, then they have an approved solution from us that they can use, meaning that when we come to a customer's Amazon Connect, it's easier for them to start using our solution, OpenQuestion, on top of that platform. That's very strong for us and very important to us. Azure is a very important partner to us, but they're not leading in the contact center space.
The leader right now is AWS, number two is Google, that's why we also needed these this year. Great references, three great ecosystems to be part of. Then we have partners that are actually taking this to market as well. Some really, really great companies out there working with us and, of course, integrating GPT, CLU, and other technologies into the solutions at the customer. A lot of our pipeline resides together with these partners. Tech Mahindra, for example, great partner with lots of great cases going on in the US as we speak. Then I'm going to go to the bottom right, and talk about some of these numbers. We already spoke about the languages, this as software as a service, API call growth is quite astounding.
The interesting thing with this is that we don't need to do anything with our organization or change anything in the platform for this growth. The way the platform has been built over the years just means it scales very, very well, which is not a given in the Software as a Service space, as we all know. Very, very strong growth, and very happy that this that the platform is supporting this growth without any problems whatsoever. 8.5 API calls happen per second in the platform during Q2. That is very strong. That also, again, is a testament to the performance of the solution, and there's no competition that's really close to that. We have some competitive solutions out there, but they're more in the space of maybe 1 API call per second in, in installations.
Huge difference in the amount of performance that we can crank out in the platform. We have our ARR target, SEK 200 million, and of course, cash flow positive before that in 2024. We are trending on that line. Our revenues are shrinking the cost gap as we speak. We've also done some cost optimizations, and we're continuing to do cost optimization. Gap is decreasing from two sides, and we're gonna be cash flow positive by the end of next year. The SEK 200 million ARR target is, of course, very important for the valuation of the company, and way beyond what any of the other high-value startups in the US that we've seen in the last few months have. Interesting stat is that Swisscom measured the transactional NPS.
Transactional NPS is the net promoter score, the customer satisfaction, how likely are you to recommend this solution to your friends, your family, your customers, the people you interact with, versus the other solution? The other solution was a human picking up the phone, and our solution was the bot picking up the phone and routing you to the right place, and it was 18 point higher transaction NPS. Lower cost and higher customer satisfaction when you implement OpenQuestion. We also have both SOC 2 and ISO 27001 certifications now, which is very important to our customer base. That means that they can entrust us with their data.
We also even support confidential compute, which we're absolutely by far the only one to do in this space, which means that it, the data is encrypted even as it's being processed, so even in memory, quite important. Now, the bottom left, look at this chart. We're really, really trending very well on this. The SaaS product was introduced in Q1 of 2021, and the customers that are now growing are the customers that we took to in Q2, Q3, Q4. We did a consolidation. We worked very hard to make sure these customers were successes and make them references, and then this year, we launched OpenQuestion to start then taking more new customers. All this growth is from that customer base, and we're very happy that they're happy with our product, but also the fact that it's growing very well.
We're exiting Q2 here with 32.1 ARR in the software-as-a-service space. You see the total revenue on the side there, approaching SEK 40 million, which is also then very, very strong numbers that we're doing right now. Very happy with the quarter, very happy with the development so far this year. Moving over to the market that we're addressing, not gonna go through all of this. Some of you have seen it before. I do wanna highlight the 48% CAGR. This is a Gartner estimate. I also wanna highlight that the market is quite small still in 2023. It's still a lot of professional services in this market, so a lot of building.
As we move along, it's gonna be more and more software content, and we are very strong in this market. We have in the range of 15%-20%. We've grown even more since we looked at this last. We're growing faster than the market. We are 15%-20% of the API calls or sessions or transactions in this space, what Gartner calls interactions, which is very, very strong, and again, the biggest references by far in this market. We have some interesting drivers. There's a law in Spain that says that a call center has to limit their wait time to three minutes. That's gonna be a driver for us. We think more countries in EU might be moving in the same direction. We're sick and tired of waiting on the phone.
We're sick and tired of pressing keypads and getting lost in long menus. We wanna get service immediately, which is, of course, what we should expect as consumers. At the top there, it says, OpenAI. Of course, ChatGPT or GPT, as is the base model here, has done a lot to this market. There's been a lot of experimentation during Q1 and Q2 by customers, a lot of people are trying. What's really being deployed right now is mostly GPT in tech companies as a part of their offering. Somebody like Grammarly, you might have seen, have deployed quite a good GPT use case.
In the enterprise so far, we have not deployed any live GPT deployments because there's still questions on how to use this, how to train the model or not to train the model, and use what's called prompt injection instead, and nobody really knows where that's gonna move. That, though, has died down a bit, the hype, and now the hype is get real use cases in, and the call center is gonna be one of those hot use cases during the autumn. Really looking forward to that, and GPT has really helped drive the awareness about what AI can do, so we're very happy with that. Here's our offering, and as you can see at the top, it says Teneo.ai. We don't take artificialsolutions.com to the market right now. That's our investor name.
What we're really taking to the market, because we saw what customers really know about, is Teneo. Teneo is the platform you open up. It's the platform you work in. It's what customers see every day as they work in, in our, in our technology, in our SaaS platform. Teneo .ai has launched a product. That's the first product we launched, and it's a way to, within 60 days, accomplish something that customers have not been able to accomplish before. That is, customers are tired of being forced to listen to what key to press. They're tired of, "Press 1 for invoice, press 2 for support, press 3 for..." Once you press 3 for 2 for support, "Do you want support on your iPad?
Do you want support on your..." The whole menu structure and tree is something that is replaced by OpenQuestion. Why do you want to do this? Well, because 15%-20% of all calls are misrouted, either because customers are just pushing their way out of the tree or because they don't understand the tree, because the tree is not their problem, it's your organization. They don't know how to maneuver these choices. We've done a big study together with ContactBabel, who have looked at this, and 15%-20% of the calls are misrouted. We can decrease that and actually give a cost saving of 9% out of the box in 60 days with OpenQuestion, and that's what we're taking to market.
Instead of a keypad, navigation, Teneo or OpenQuestion, the solution based on Teneo, picks up the phone, listens to why you're calling today, you express yourself in a natural sentence, and it will intelligently route you to the right person, and that's the first step. After that, you can build all sorts of automations and do all sorts of things to make customers even happier, but that's the first step, and it takes 60 days to deploy. It used to take years to deploy this type of solution. Results are happier customers and lower total cost of ownership. We have the references to prove this. Again, the only ones in the space to have large references. Swisscom, very big user of our system. Telefónica, a very big user of the system, and they're happy to talk about this.
There's YouTube videos, Vimeo videos, all sorts of testimonials, and of course, other large users of this platform are very happy in using it. Biggest reference, this, in a space that's growing, in a space that's got more attention now with AI and conversational AI focus, thanks to GPT, and that's what we're building our pipeline on to get the new clients in now that we lowered the cost of bringing new clients in. Very happy with the development, with the pipe, and with this solution. Just very slow, just a brief summary of where we stand and why we are where we are today. We started in the year 2000, started building on technology. It was what we call a professional services-driven black box. We would install it, we would run it for the customer.
That's what most conversational AI companies do today. We then started with what we call fake SaaS, i.e., we put something in the customer site. Sorry, we put it in our own server, our own data center, but we ran it for the customer, but it was still a server or a virtual server for that customer, still supported by professional services. We started doing omnichannel and transactability very far ahead. We then took that technology, patented it, rebuilt it into a software as a service platform, where everybody runs in the same code base. Of course, they're all separated, so there's no data leakage, and that's proven again by certifications. It means that we can now run more customers with the same organization, instead of having to scale the organization with customers.
Very strong SaaS offering built in 2021. Started getting our first customers, started getting the feel for what solution to develop on that platform. That's Open Question. Also how to run that platform in a cost-efficient way, because it was quite costly to run it in the beginning, and during this autumn, that cost will go down, and that enables us to bring in more customers as well in this, and that's the running of the complete platform. Open Question launched during 2022. Great references. In 2023, we wanted the partnering with Google and Amazon, as well as the Microsoft that we already had, get on their marketplaces, and we started then pushing Open Question in the market, end of Q1, and are now looking for sales expansion in the US and some countries, the large countries in Europe.
That's our, that's our focus space. That's our brief history. It took some time to get here, but we are certainly poised for growth ahead, and we're already delivering growth as well. Here's some July numbers that we're sharing today, since this report is a bit late. Of course, being a small company, we're waiting for all the big ones to report first. It being a bit late, we also have the benefit of having the July numbers already. The SaaS ARR growing from July 2022 to July 2023, from SEK 21 million-SEK 34 million, so massively outperforming the market still, and the API call volume growing very healthily, and more than doubling, which is also a testament then to customers really growing their usage of this platform.
Of course, API calls makes the biggest gross margin, but customers are expanding seats and the base platform, which is why the revenue is expanding a bit slower than the API calls, but we will see more connection between that in the future as the gap, as the API calls become a larger part of the total revenue. I'm gonna give it over to our Chief Financial Officer, Fredrik, to talk a bit about the Q2 operations and numbers.
Switch on here. Good. Let's start a bit. We have some really promising highlights that we also kind of showcased in, in, in our Q2 report. We now are partnering with the Google, and that is a very important step in us, you know, being on the three hyperscalers platforms. Also we have, we are now available also on Genesys, Genesys AppFoundry, with our OpenQuestion solution.
In addition, we also held an AGM in June, and the key highlights from the AGM was that we now have 2 new board members with Lars Rinnan and Mathias Björkman joining the board, and both of them with a very knowledgeable in the SaaS space and also in the contact center software space. Very happy to have them on board. We also had a new warrant program launched. Also what can be important for shareholders to acknowledge there is also that the existing employees in order to take part of the new warrant program, they also have agreed to surrender their old warrants in the previous warrant program.
Those warrants will be canceled here in Q3, when we have onboarded all of the employees in the new warrant program. Some of the old warrants will be canceled, basically. Then, as Per mentioned, I mean, we have very, very strong growth in our SaaS ARR, 60% plus year-over-year, June versus June. Very satisfying numbers there, and also the same then also on API call volumes. I will go into more details on EBITDA and OpEx going forward, but the kind of trimming activities that we have taken on reducing OpEx is, yeah, is clearly visible in our numbers in Q2 and will also be perhaps a bit more visible also in Q3 as well.
Per, maybe we can move to the next slide. Thank you. On this slide, I mean, I think I will just mention a few, few highlights. I think we, we have mentioned a lot already on our SaaS ARR and API call development. Obviously, that we have, you know, the sales metrics that keep growing, both on our SaaS business, but also on our on-prem business is really satisfying. Of course, we are also then experiencing strong quarter-over-quarter and year-over-year sales growth on basically all sales metrics, whether that is on SaaS or on on-prem. Really happy about that. Our recurring software revenues amounted to SEK 14.6 million in Q2, plus 60% versus Q2 2022.
Obviously, we are then also growing faster on our recurring revenues than on total net sales. That is also a consequence of the transitioning from a kind of hybrid PS license model into a pure software model or SaaS model. Our net sales amounted to SEK 14.8 million in Q2 2023, versus SEK 10.2 million in Q2 2022, so up 45%. Basically, the delta then SEK 200K, roughly in non-recurring revenues only. That is also, I mean, if we look on the graph, we can also see that we now in the quarter had 99% recurring revenues as percentage of total sales.
That is, of course, an important metric in showcasing we are now a pure software company today. Q2 gross margins amounted to 72% and was a big improvement versus Q1, partially because of lower commission, but also that we also increased our volumes on both the SaaS volumes, but also on the on-prem volumes. That's a consequence of that. We also continued to reduce our OpEx in line with what we guided in our Q1 report. Headcount-wise, I mean, we are, we are 8 people less when we leave Q2 versus when we left Q1 this year.
That is obviously being visible in numbers, but we have also on top of that also, of course, run through other OpEx as well to become more efficient in how we run our business. This also resulted in improving our EBITDA. Our EBITDA, adjusted EBITDA amounted to SEK 13.9 million in this quarter, versus close to SEK 21 million Q2 last year. A significant improvement, and I will come back a bit more on OpEx as well on a coming slide. We also ended the quarter with close to SEK 60 million in the bank. I should also say that in this number, we have one customer, a big customer who has changed a bit their way of invoicing.
I think if we would have received that, that invoice, we would have more or less close to $60 million in, in the bank, end of Q2 2023. On solid grounds for, for continued growth, basically. We can go to the next slide, please, Per. Yep. Just briefly on, on our SaaS API call volumes. Per mentioned this already, but this is obviously a, a, a key KPI or indicator on how our customers' application, and, and the usage of them, and also as a, as a trend follower in, in showcasing how they are growing, whether they are growing or not growing. And basically, you can say the more applications/solutions and covered regions, /use case, number of use cases, the higher API call volumes.
Our SaaS API call volumes are growing in a fantastic way, and the growth pace in, in July 2023 is even higher than in June 2023. We can also see that, I mean, those two months are, you know, record all-time high months in terms of, you know, volumes and revenues. We also expect that our existing customers will continue to grow volumes. The key driver for that is, is also additional live use cases that we expect to see volumes from in the second half of 2023.
And then one interesting note on, on this is also that, given the volumes we have now on the SaaS API, on SaaS API calls, we can just see that if we are continuing to grow faster than the market, so i.e., 48%, according to Gartner, our SaaS API call volumes would then, I mean, year-end, this year, more or less have 10, grown 10 times in 2 years' time. I think that is quite an amazing number just to see that we, we have actually delivered a lot during these 2 years.
If we look on concrete numbers, we can just see that the monthly SaaS API call volumes grew 117% year-over-year, June 2023, and also then 7% quarter-over-quarter. If we look ahead into July, we can also see that July 2023 versus June 2023 showed an increase of 11%. Growth continues also following end of quarter. I think we can go to the next slide, Per. Yep. Not to say too much about the obvious here, of course, the growth in our SaaS model is also evident in our other key metrics, SaaS ARR.
SaaS ARR grew with 60% year-over-year, June 2022 versus June 2023, and 61% year-over-year, July 2023 versus July 2022. Then we also saw a month-over-month growth, July versus June, of 7%. Same message as on, on API call volumes. We see, yeah, very strong growth, and going forward, it's looking very promising on continuing, growing volumes as well. Next slide, Per. Yep. Yeah. This is basically, I mean, just outlying both the SaaS business and also the, the, the on-prem business. Obviously, as you can see from, from, the right-hand side there, total ARR growing 57% year-over-year, and SaaS ARR growing at 60% year-over-year.
Clearly, we see growth, a bit higher growth, on SaaS, but also very, very solid numbers also coming from, from the on-prem business part of, of the, our total business. Of course, I mean, as I already mentioned, I mean, the transitioning into a pure software company, it's very encouraging, to see that our recurring revenues is, is continuing to, to grow in, in a very high pace. Currently, we do see significant, volume potential also from our existing customers to continue to grow. Yeah, we are very confident in, in, in the continued growth, growth, basically. Overall, I mean, just to conclude, a very strong performance in, in the quarter.
Also if we look versus also with our internal budgeting, I think we're exceeding those numbers quite a lot year to date. I think it looks very promising going forward. Per, OpEx slide, please. Just as we mentioned in Q1, we have simplified our operations, and as part of that, we have also trimmed the organization and also reduced the number of headcounts with 8 people quarter-over-quarter, but also reduced other OpEx costs in the quarter, and we see the impact in reduced OpEx already in the quarter.
We have also provided a bit, kind of, adjusted numbers on this as well as, as there are some costs that, or activities that we have implemented and we know will be visible in Q3 as well. On this, we are not planning any further headcount increases nor decreases short term. Now with annual adjusted OpEx run rate going from $116 million in Q1 to $113 million in this quarter, and that is obviously excluding extraordinary items and cost of sales. All in all, this also means that we currently have a monthly OpEx level of slightly more than $9 million.
Also if we look for the rest of the year as to, to give some guidance, I think we will obviously continue to trim OpEx, but also our costs of sales, costs, which is the, which is also part of, of our compute cost, for delivering the, the SaaS, the SaaS, business to, to our customers. We will continue to trim those costs, but we, we do not see any material reduction from, from the current OpEx levels. I will conclude with that. Next, over to you, Per. Yep.
There we go. We're sitting in the same room, so we have to mute. Okay, so, I'm gonna leave this slide up while speaking a bit about what's what I see in the market right now. The, the pace of innovation in the models is very high. The pace of innovation in terms of using this for enterprise is not as high. We have the references to use this AI technology. We have the partnerships, we have the technology, but we also slightly different than the startups. The startups in the U.S., which are now being valued at, you know, crazy money. We just saw LangChain, for example, being valued at 200x earnings or 200x ARR. It's a, it's a very different world out there. Yes, we have debt.
We have a great, a weird debt from Capital Four, great partner to us, that we need to repay to them in 2026. It has no cash impact until then, and is a very stable thing, and it's gonna be repayable with our business. Yes, we are traded on the... We're publicly traded. Yes, we've been around for a while. Yes, we're still cash flow negative, but we're shrinking that. I just think this, this, what we're delivering is absolutely fantastic in the AI space right now. With partnerships with the three hyperscalers, I think that we're poised for some really, really nice growth the coming years. Very happy with the quarter, but very happy where, where we stand as well as a company. I wanna now move over to taking some questions.
I'm gonna shut down that PowerPoint and see if there's any questions out there. Now I forgot the instructions. It's a double thing, right? You have to unmute twice, I believe, to ask a question. Actually, no, I have to unmute you as well. Is that the way it works? If you raise your hand, I will unmute you, and then you unmute you as well. That's the way it works. Forbes raised his hand. I'm gonna find him in the list. I shall unmute Forbes. There we go. Now you need to unmute yourself again, Forbes.
Yes. Hi, can you hear me well?
Yes.
Great. Yes, good morning, and thank you for the presentation. A couple of questions on your customers, I'll take them one by one. No new customers in this quarter, but you say that you're building a very strong pipeline. Can you tell us a little bit about this and which sort of customers that you're engaging with?
The pipeline is all large contact center enterprise players, and it's, it's together with these partners and hyperscalers. It's customers that have, for example, a Genesys platform set up. They might have a Nuance IVR, they have keypad navigation, and they're looking to change the experience for the customer. We are one part of that, so OpenQuestion is one part of that. The interesting thing with OpenQuestion is we can stand it up in parallel to Nuance, for example, and then run it. In 60 days, you'll be able to shift your load over to that. That's the, the customer space we're in. It is an enterprise, which means it's a procurement process.
Even if we are at the stage of being in legal, that means that the last stage of the sales process, that's where we see a procurement being connected in, of course. Even though we're flexible in our terms, in terms of not having to commit to 12 months, procurement understands that once it's in, it's gonna be costly over month, over month. We do get into a procurement process, which takes a while, but, yeah, we're at late stage with several of these opportunities going into the autumn. Quite, quite sure that we'll see at least 2 or 3 large enterprises on the customer list during Q- Q3 and Q4.
Right. Understood. Thanks. Also, it looks like your legacy customers are growing at a very satisfactory rate. You've previously spoken about transferring all of these to SaaS at one point. Could you give us the latest on that?
The latest is we have 1 customer who is basically saying that it's not gonna happen. But the other customers, we're now in the, in the stage where we've ticked the technical boxes, and that was the important thing, was ISO and SOC, to be able to do this. We believe we'll be able to move all but 1 over to software as a service within Q1 of next year, which is, has to do with when their contracts actually expire. We, we're now done the technical stuff, now it's the commercial stuff. It's a change of how they pay. Some of these have some very favorable terms, and of course, we're, we're gonna be keeping that $0.008 per API call.
That's, that's what's still outstanding, except for one, where we basically can't really see how we're gonna be able to solve it.
Perfect. A final question from me, and it's on the July numbers. The, the growth in July over June was, really strong, and I mean, it looks that it's, even stronger than all of the growth in Q2. Are there any particular events that we should be aware of that, that were kind of driving that?
I think the events were more in Q2. If we look at the numbers, we had so we now have more than 50% of the revenue in the US. In the US, they had spring break, so that, that took down the numbers, but also Easter took down the numbers. We had several things in, in Q2 that were bringing down the numbers in Q2. I think that's more that than it is something in Q3. 'Cause Q3 is also typically a vacation, and July is typically also a vacation month, but of course not in to any great extent in the US, except for Fourth of July. We think that that's seasonality versus the fact that Q2 has a lot of short, short weeks.
Mm. Perfect. Okay, well, thanks. I'll, I'll go back, back to the queue. Thanks.
Thanks, Forbes. We have Viktor raise his hand. I'm gonna unmute you, Viktor, and then you need to unmute you as well.
Can you hear me now?
We hear you.
Yeah, great. One follow-up here on the customer side. You, you mentioned in the report that you canceled the agreement with the outbound customer, SelectQuote, due to some, some high risks. Could you give some more color on, on that one?
The way that was set up, we felt that there was a major risk of being ensnared in this. There's this framework of laws which are different state by state, in how outbound calling is, can be done. Being a software as a service, we felt that we would be, could become legally liable if this was breached. The fines were just too high to be able to accept that business risk. We're looking at a different commercial arrangement. The technology still works. We're looking to see if we can do a different commercial engagement there, but we haven't really found one that sort of protects us against the risk. It's similar to. Actually, the fines are similar to the risks that we run if we were to engage in real healthcare.
Not the administrative part of healthcare, but the real healthcare, which is also why we're not in that space. It's slightly surprising, but that's the, the advice we got at the end from, from our legal partner. We'll see if we can find a different way of structuring that.
Yeah. All right, understood. Also, I mean, how would you say the competitive landscape has, has changed in the last couple of months now with the, with the high interest-
So-
-of AI solutions?
Right. Essentially, we thought that maybe ChatGPT would drive use cases around chat, because that's how the CEOs first saw the GPT model. We also see that the contact center is really becoming a large portion of this. The competitive landscape has changed in a bit in that the big consulting houses, that being Accenture and Ernst & Young, et cetera, they are building solutions for the contact center on GPT, but those are all trials, and they're not gonna be able to scale the way we do. It is changing a bit in that we now again have professional services as a bit of competition rather than ready-made solutions. We really see that our offering together with GPT is something that customers are willing to listen to instead of that PS...
We think, we think we have a strong offer in that. It has changed a bit in that the PS companies have gained more ground. Also, of course, because recruiting people that understand the space is kind of difficult. What large enterprises do is they go out and, and procure that knowledge from, from these large consulting houses, and they, of course, like to bill that as hourly billing because it's very high or day, day rates. The day rates are very high. Right now we see, you know, in the $5,000 a day type rates right now for people in the LLM space. Slight, slight change, less focus on things like... What you could say also, though, the one we've replaced the most is Nuance.
If you look at what customers typically have, is they have a Nuance-based IVR. Microsoft has done a very big defocus from Nuance, given the big focus that they have on, on GPT. We saw during, at the end of their financial year, that Nuance has been. There's a lot fewer people working at Nuance. I don't know how to express that in a good way, but if you look at the, the, the announcements, Nuance has really been defocused. There's less solution, push in that space. We're competing against the PS people, but we believe we have the right positioning there. It's a long answer, but it's, it's an exciting space.
Yeah, thank you.
happening at the same time.
Thank you. That was all for me.
Thanks. Now, we have Niklas. I'm gonna unmute you, Niklas, and then you go ahead and unmute yourself. There we go.
Okay. Can you hear me now? Good morning.
We hear you. We hear you.
Excellent. Good morning, Per and Fredrik. Let's start off, thank you for a great presentation. always interesting to listen to, and, I have the greatest respect for the, the job you're doing. Must be very challenging to sort of, build a new market in a way. having said that, I'm maybe somewhat worried. look, you know, good, perhaps even great. I'm worried that... I find it a little bit challenging to understand that it takes such a long time to convert, potential clients to new clients. given your plus 20-year history, you have the world's most well-known and reputable sort of partners, and you have blue-chip clients, and you have a relatively, now at least, solid cash position, and still, it's, it's proven very challenging to, to gain new clients. that's part of the question.
Then adding to that, I am worried you're not trimming your OpEx sufficiently. Instead of personnel costs kind of decreasing by SEK 6 million, it's actually increasing by SEK 6 million, which I find also awkward, given that you are trimming the organization. I would have personally liked to see, and obviously I am a minor investor with limited knowledge of your business, that the personnel costs were decreased by SEK 6 million as opposed to the opposite. Are you confident that you are trimming the organization enough so as to avoid a new rights offering before you turn cash flow positive end of 2024? That is my first question.
Right. Let me start from the, the, the top, the new customers. That is obviously a big focus of ours, to get new logos in on the platform, since the platform can scale, and that's, of course, gonna give us gross margin. What, what, what's happened in the industry is that what we used to sell Teneo is in what's called conversational AI. Conversational AI has become something that every man and his dog sells. Now, very few of every man and his dog actually have references, but we have a, a good case here in Sweden, where a partner of ours, we do not transact in Sweden, but a partner of ours bid a very good solution to a customer that actually has almost 1 million phone calls, a month.
Mm.
The, the solution that the partner was bidding was, together with the, the services and our solution, the customer thought that they would pay in three years about SEK 10 million for that solution.
Mm.
They, they got an offer from a fairly reputable PS company in Sweden to do the same for SEK 300 thousand over three years.
Mm.
The difference is, they would use open-source products and try to, you know, build that together.
Mm.
That's not gonna work, it's very difficult to compete in the procurement process because you are constantly have to lower your price. We decided-
Okay. Fair enough.
to not compete in that process.
Mm.
We stopper Teneo as a platform because we don't wanna compete in that. That's sort of Q3, Q4 of last year, because the competition just became too high. One reason it became too high is that Microsoft, who is positioning our platform, they retired their platform Composer and made it open source. That platform Composer could look like something that could be usable, but we also know that Composer never really delivered any large use cases, and that's why Microsoft retired it. It's open source, so you can build on that. That happened in Q3, Q4 of last year. That's why we went OpenQuestion.
That's why from a pipeline perspective, even the conversational AI pipeline, where we have great reference stories, so we have, you know, somebody in Germany who has met with Swisscom or Telefónica Germany, even there, we see that even if we still have that pipeline, we're probably not gonna try to convert it because it's just gonna demand- it's a race to the bottom in pricing. OpenQuestion demands the performance that we have, so you can't deliver OpenQuestion, and because we do it in 60 days, it's no longer a platform, it's no longer a something you have to build, it's a solution, it's a product, and that, that just changes the pricing dynamic. You save 8% on your call center staff, you pay us 0.008 per outgoing call. It's a very simple conversation. That's the repositioning.
We're really rebuilding the pipeline to be on OpenQuestion rather than conversational AI, and that's, that's the timing factor here. In that space, we needed those three partnerships because Amazon and Google are winning that space. That's, that's the first question.
Mm.
Second question around cost. One part of what Fredrik was talking about, reducing the total cost, is that we're bringing some services in-house, which is increasing the staff cost a bit, but that's gonna reduce, reduce the cost a lot and increase the gross margin a lot, because we have some cost that scale with customer acquisition, due to the fact that we have external management of our cloud service, parts of our cloud service today. So think the 24/7 type services, those are done by an external party. So part of that, bringing the cost up on, on staff is gonna bring down the cost on, on the rest. Then for in terms of rights issue or not rights issue, that's totally up to the board.
Of course, we are tracking on that revenue trajectory that we, that we've made the plans on. From that perspective, we're closing the gap the way we wanted to close the gap and the way that we've been forecasting to close the gap. We're still tracking towards that SEK 200 million ARR. Yeah, I can't really speak about... It's not, they wouldn't, they wouldn't ask me. I was, maybe that's the way to put it.
Right
owner and board discussion. Of course, the AGM did say that, you probably saw that, the AGM did approve some rights issue as well.
Okay, that, well, final question then. I mean, obviously, personally very happy if you, you earn a decent enough, enough money or amount of money on your warrants eventually. It seems to me that over the past couple of years, you sort of halved the price, the strike price for the warrants. Understandable, and now it's more than half of, you know, from 8 going down to 3.2. It seems now you're setting the bar not relatively low, 3.2 into 2026. I mean, again, personally happy that for you and the team doing a great job, it should be there, but isn't that? Yeah, how should I say? That's that seems to be more than achievable, sort of 3 years from now.
I guess what I'm asking is now, now it's, it seems rather low, or, or how did you come up with 3.2?
Well, that's set in relation to the stock price at this point, right? The, the, the, the warrants.
Fair enough.
... and, and, our, our, perspective or my perspective on the warrants is that without the warrants, it's very difficult to do in this space. If people can't see that they're gonna make money on that, it's gonna be very difficult.
Mm
... to retain staff. One thing that GPT.
Right
as, as we were talking with Viktor about competition, there is competition for talent.
Yes.
And-
Fair enough.
... I just, I just read in The Wall Street Journal last Friday that Google is now paying $1 million for an AI engineer. Now, that's if you're willing to live in Palo Alto, which of course is, you probably pay $1 million-
Yes
... in rent.
Right.
still-
Yeah
... the competition is quite hard, so we need to make sure that our staff-
Sure
... have a, a reasonable
Yes
reasonable chance-
Fair
... of making money on those as well. We did kill all the old options, so from a dilution perspective...
Yes
at least that helps.
Yeah. Thank you, Per, Fredrik. You're doing a great job.
Thank you, Niklas.
Appreciate it.
Are there any other questions at this point? I'll give it. Nope, seems to be no more questions. Thank you all again. Great quarter, exciting area, and then now is the time to start delivering some new customers as well, like you said, Niklas. Thank you all. Thank you very much.
Thank you.
Bye.
Bye.