Okay, I'm going to go ahead and start. Fredrik, a thumbs up if the sound is working. Excellent. So good morning, all. The presentation is, as always, available on investors.artificialsolutions.com. Just as a backgrounder, we are going to market on Teneo.ai. So Teneo.ai is our go-to-market site, and Artificial Solutions is, of course, the name of the company that runs Teneo.ai. And for our investor site and our public and traded stock is ASAI. You'll find the presentation on investors.artificialsolutions.com. This is being recorded, just so you all know. And I'm gonna kick off this with today, a bit of a different layout. We're gonna do a business update just to get back to a bit of what we are doing.
There's been some very interesting development during Q4 on the business side, which, which I think is important to highlight. We're gonna go through that, and then move into the numbers, and then I go to Q&A as usual. To start off with, it's Per and Fredrik representing. We've now been here for the better part of three years. And, I come from a background, together with a large portion of the other management team, the background from a company called IPsoft Amelia, which is basically in the same space, the same technology space, a different go-to-market space as Artificial Solutions. And Fredrik, a long background in different PEs, but also has worked in as a CFO in software as a service company before coming to us.
We've done a bit of a turnaround of this company in terms of strategy, and we'll come into a bit where we are on that journey. From a business update perspective, so our vision is that we, we can experience the world without queues, without keypad navigation, with instant service, and we are delivering on that now. And also what we have, which I'm gonna show today, is that we also now have real proof on why it is that we can deliver that better than anybody else in the world. And I do mean anybody else. That includes Google, IBM, Amazon, or whoever else you might put into that category. What we do is we reduce misrouting in call centers, and we also automate call centers.
So many of you might have read about Klarna lately, which is a, a chatbot that responds to, to, requests into the contact center or the customer care center. We do similar things at our customers, except of course, our customers are older and much more complex environments, and also they use voice, and that's where really the differentiation comes in. When you use voice, it becomes very, very difficult, and then what's stated here, the 99% NLU accuracy becomes incredibly important. I'll come back to that number, 'cause that is quite, quite an important number. All this gives money to our customers.
Our customers are realizing great gains by reducing the amount of people in the call center, but also by having better transactional NPS scores, so better scores of, "Would you recommend my company to another company after having talked to the customer care bot?" So where are we on this journey that I talked about? It started out with a new management team back in 2021. We then started with rebuilding the product to a SaaS product. That was a rather large re-architecting of the environment, but then was launched as Teneo.ai. Again, that's our go-to-market. And we also divested professional services. So the company had revenues from implementing, which we no longer have in our books, so that's also a large deduction in revenue at that point.
And then in the beginning of 2022, so we sold to some large customers in 2021. Beginning of 2022, we took these first customers live, and started seeing what they were using the technology for. So this new SaaS platform, so that is something we host for the customer, and the customer logs in and builds their solution in that, or a partner builds that solution for them. We saw that what they were using this for was what we call Open Question, and have now productized during 2022. And the Open Question is essentially, pick up the phone, understand why somebody's calling, and to a large extent, automate that phone call without the need for a human, so make that into an, what's called an unattended call. We then built the reference stories around this, published those.
We focused on the growth in existing customers to make sure that they all became large users of this technology, and that's what we've been doing during 2023. We also set up with tech partnering with the three cloud providers, the three big ecosystems that are out there today. Number one, with Azure, and that's actually where I am today, so we're meeting up with the Microsoft team here in Seattle. And then we have GCP and AWS. Slightly different tunes on these partnerships. I'll get in a bit more into them in a later slide, but very important for us in this space, since everybody uses these now to in different parts of their customer journey, they use one or maybe two of these ecosystems.
Then we also started partnering with implementers, so the people that take this to market, that help with the staff transformation, but also the tech transformation, largest one being Tech Mahindra and CGI and Valcon. And then the new logo hunting, we started, let's say, tail end of last year. And also this year, the big focus is to take us to a cash flow positive month, which is a goal that we shared two years ago. Okay, so partnerships with hyperscalers that I mentioned, let me go into that a bit more. On the Azure side, we are mostly partnering up on the AI services side, and that's also who the people we meet here. That's mostly for our product and our development.
So that means that we, for example, were part of using GPT or what's called Azure OpenAI early on in our development. And we've been using that, and we're talking about two years, and also launched products during this quarter, the Q4 quarter. We launched some products in that space, which I'll come back to a bit, too, as well. Google, mostly just that you have to be partnered with Google, and we are, we are a partner with Google. We have set up that we are in their marketplace, and we're an approved vendor. The interesting thing with them is that they are not necessarily getting into the contact center, but a lot of customers are still using some of their technology.
So we did our first deal in the Google space together at HelloFresh during August of this year. August, or August of last year. But that is not a go-to-market partnership. AWS is more of a go-to-market partnership. In the AWS space, we are targeting the customers that are buying Amazon Connect, which is their big contact center platform. So they managed to break in and actually selling large contact center platforms here. Contact center, you could just translate that with a platform that manages all the phone calls and all the agents that pick up the phone calls and all the automations in that piece of software, which used to be something you bought from Ericsson and put in a big room, which nowadays, of course, is cloud-based. So all of that, we take to market in Teneo.ai.
We are selling this in the Azure, what's called transact model, meaning that we don't have to negotiate with these large companies, to on the terms and conditions of the contract. They can draw this down and make this part of what they call their MACC agreement, their large agreement with Microsoft. So quite an important way for us to work on the deals that we now have in pipeline, which make at least a part of that deal process easier, which would be the legal part. So what this does is it provides us access to the blue-chip enterprise customers. You need to be able to tackle all these three to be relevant for somebody like AT&T and Swisscom. And as you see here, these customers are talking about us.
We're talking about at AT&T, it's Smart Home, which is anything that has to do with the home. It's fiber, it's routers, it's alarm systems, et cetera. They have a much higher customer recovery rate, so meaning that they don't have to send somebody out to the customer. The customer talks to the system and resolves 15% of those issues without having to send a person out to the actual home of the customer. That's a large gain for them, of course, as that's quite expensive to do. Swisscom have also measured their customer inquiry hotline, so when you call, let's say, 1-800 Swisscom, and they're considerably more productive, according to their CEO.
Also, they've done testing on the transaction NPS and found that it's higher when the bot picks up the phone and resolves immediately, versus a traditional approach, where you come to the queue and wait for a human and then get your solution. So we have a unique relationship with Microsoft in the AI space. We're part of what's called the AI Inner Circle, which helps a lot with these products that we also launched during Q4. And the last portion of this at the bottom is quite important, and I'm also going to go into how that can be. But we are the only player today that answers millions of phone calls per month in several customers. We have several customers that are above a million phone calls per month.
We even have one which has a million automated, so unattended, so calls that no human need to take, and I have an example of that moving into this as well. So it is truly a unique solution in the market today, which, of course, is what we are now taking to market with the references and with the stories, and now also the proof points, which I'm gonna show to you. So what I'm gonna do now is I'm gonna show you. I don't know how if you read Telekom idag, but I definitely get those emails, and I saw that Tele2 won Customer Care of the Year in 2023.
This is a recording of a call into Tele2, and I'm gonna put you through the pain of listening to this for two hours, and I'm gonna ask Fredrik to do a thumbs up if the sound works. I actually think it's not gonna work. I need to take off my AirPods. Two seconds. Now, can you hear me, Fredrik? Thumbs up. Okay, now I'm gonna try the sound and see if it works. Did the sound work, Fredrik? Let's try again. Teams doesn't wanna do it when you have your AirPods plugged in, so now my AirPods are out. Let's see if that works.
Welcome to Tele2 customer service for private customers. Are you a business customer? Please contact us at 94, 4, 4. For English, press star. Please enter your social security number with 10 digits and with hash. You will now be given a number of options. Please listen carefully to all the options before selecting the correct one for your inquiry. For technical support and troubleshooting, press one. Invoice questions, press two. Buy services or products, press three. Terminate or cancel subscriptions, press four. Other information about services or changing subscriptions, press five. For your and your safety, we would like you to identify yourself through BankID. Would you like to use BankID for identification? Press one. If you don't want to or are not able to identify yourself with BankID, please press two.
So note how, first of all, you've all seen this, of course, I'm not expecting any reactions on this team's call, but, you've all suffered through a keypad navigation. But also notice how disjointed it is. So it goes back to Swedish, although it's English, because it's all scripted, and it's a very complex tree that they have to build. And also because it's complex to do identification in a, in this setting, they do it upfront, even though if it's a support question, you may not need to identify yourself, so that's a waste of time. So that was 2 minutes and 30 seconds, and you still haven't solved anything.
You may have pressed the right button, and in fact, ContactBabel , which researches this, says that in 12% of the cases, enterprises, people do not push the right button, meaning they end up in the wrong queue. So now they've done this, they come to the queue, and then they need to be rerouted, which both costs money, but of course, frustration. So people are quite frustrated by the time they actually get to the agent to talk to, as you can see here on the picture. Now, contrast this with the alternative solution, which is the one that's deployed here at Microsoft, here in Seattle. It's 1-800 Microsoft. Obviously, that's not the number, but they have several different numbers, and this is the way you talk about customer care in the US. You say 1-800.
And then, this, in this case, same thing, we're calling in the support to get support of a technical issue. This time, Teneo picks up the phone. Note how they do it all in a way that is good for the customer. So, for example, they all the regulatory stuff, the stuff about being recorded and so forth, all that they do in a speed, which is 2.7 times the normal speaking speed. 'Cause if you don't hear that, just hang up and call again. It's not a big deal. But that's most people that have called the call center have already heard all that, so they do that faster. But most importantly, it's an OpenQuestion format, meaning you don't, you're not dictating what the customer can and cannot respond.
You can then also see that this is a fully automated call. I'm gonna cut it towards the end, but the system has now understood why the customer is calling, and we can establish what the next action is.
Hi, thanks for calling Microsoft. To help us to improve the quality of our products, services, and training, this call may be recorded or monitored, and information collected on this call may be transferred to other countries. To help me best assist you, I need to know if you are calling as a home user or a business user?
A home user.
Looking forward to helping you. Please tell me what I can help with.
Yeah, I got an issue with my Surface. The screen is blank.
Let's get you some assistance with that. Please tell me a bit more about the issue you are experiencing with your device.
Yeah, when I move the mouse around, it doesn't turn on.
Okay, I can help you troubleshoot your power issue on this call. Would you like that?
Yes.
Okay, let's get started. Is your device currently plugged into a power outlet?
Yeah, it is.
Okay, next, is the tip of your charging cable lit?
Yes, it is.
Hold down the power button until you see the logo screen. This will force a restart and may take up to 20 seconds. I'll give you a chance to do that now, and we'll check back with you in a bit. Just let me know at any time when you have completed this step.
Okay, I completed it.
Okay, did this help to resolve your power issue?
Yeah, it did. Thank you.
Great! I'm glad I was able to help resolve your issue. Hi, thanks for.
In this case, the customer calls in. You don't need to identify the customer because you're just solving a problem for the customer, and the customer is, of course, quite happy with this. This happens now about a million times a month, and you can make a quick calculation on that that probably corresponds to about 1,000 call center FTEs, full-time equivalents, that do not need to be used for this process anymore, can be used for something else, or in the case, if there is, if it's BPO, they can be used at another customer, most likely. Very powerful, but also very, very high ROI on this type of solution. When we play this for customers, it's not like somebody else has something else.
Now, it's still quite complex to sell it because it's a large moving target here, because you're often changing your contact center system to that Amazon Connect or one of the other systems in the market. But we still haven't found anybody that says, "Oh, I have that already." That just doesn't exist. And we also ask contact people to go through all those 1,000 enterprise customers that they work with, and we still haven't found anything like this. So quite confident that this is a unique solution in the market today. And the difficulty of understanding natural language is we, we need a proof point of why we can succeed with that within Teneo.
What we did is we used a data set called Banking77, which is freely available on Hugging Face, which is banking conversations, so bank customer talking to a bank employee. What we did is we tested the accuracy, so we know what the customers were talking about. We used a testing tool called Cyara, which is the primary testing tool for anything natural language and used by everybody now developing LLMs and so forth. That's how you sort of track hallucinations. You can't fix the hallucinations with it, but you can flag the hallucinations with Cyara. Great product. What we did then is we tested that with different. So others have tested, for example, Google Dialogflow; they tested it themselves on Banking77.
We tested Amazon Lex and Teneo with Cyara, and here's the accuracy number. Now, I'm gonna compare this a bit to humans. So a human would be in the 0.92 range if that human is in the same country as the person. So if you're both in Bulgaria, and you both speak Bulgarian, you'd be. Oh, sorry, you'd be in the 0.92. If it's outsourced, it'd be somewhere right below Amazon Lex here, 0.88. But 0.95 in the initial dataset test is just tremendously better, and that's why this solution actually works, because you don't get the errors that you get with other solutions out there. So Google, you might remember, they launched a solution where you could call the hairdresser and book your time and rebook your time. That solution never really got to market.
And you can understand why, when you see that they only, they only pick out 74% of what the customer said, which is gonna be, of course, you're gonna be missing, was it Tuesday or was it two o'clock? So you're not gonna be able to find that, the accuracy in that type of solution. So we now know that that's the, the solution as well. What we've done is that we have now just, that was last week, presented this at the big CRM CX event, and people, of course, were ecstatic of seeing that accuracy and that that is the real reason why we can deliver something like we do at Microsoft. So we are, today, a, a prominent player, so, we power the most amount of transactions as well in this market.
It's a market that's growing at about 38% CAGR. We have the largest references in this high-growth AI market today. The market is even more in focus now because of LLM, and of course, the fact that Klarna went out and said that it actually saves FTEs is quite positive for us. None of our customers will ever do that. It's not something that large enterprises typically do, but very interesting that Klarna did that. And also, the business model is scalable, as we will see, with improving gross margins in Q4, but among these existing customers, the gross margin just keeps improving quarter-over-quarter. It's very difficult to switch this out, and that has to do with that accuracy, which has to do with the patented technology we have.
It's the way we store that natural language and process the natural language, and that also is how you can afterwards train the model and make it even better and better. So it's quite difficult, it would be quite difficult to switch us out at a customer. So we have rolling one-year contracts with the customers, which also lets us adjust the pricing, which we've done in the last two years. So we feel quite confident that customers like using the solution and that they gain much more value than the cost out of the solution. And then, of course, this platform also lets you use the absolute latest technologies like GPT, and that's what's gonna, what I'm gonna talk a bit about, the events during the quarter. Short status update on the strategic review. So this is from the board of directors.
I'm just reading it out loud, or rather, just telling you guys about this. But the board of directors announced on October 26th that they initiated a review of the alternative, the strategic alternatives, to maximize shareholder value on Teneo.ai or Artificial Solutions. And that's because of the evolution in the market landscape. Partially, the references, the attention we get in the U.S., of course, but also the fact that the LLM space, where we also have now products in use in enterprise, is quite interesting. So it's still ongoing. The board senses that there's still interested parties. There's no timetable for completion of this review, and also, the board is not committed to anything in terms of doing any further announcements on this until decisions are made by the board.
The board also emphasizes, emphasizes that this review discussion. We are a leading player in the AI and contact center software vertical, and we're recognized by many of the large players in the industry as a leader, and that's why they feel that this ongoing strategic review discussions are gonna keep on going. So again, not something that Fredrik and I are directly involved in, but let's just phrase it as there's interest for what we do, and the board is looking how to capitalize on that best way for the shareholders. Financial. Some quick summary numbers. I'll go through the events, and then I'm gonna hand it over to Fredrik to deep dive into the numbers. We had a net revenue retention of 145% in Q4 2023.
We have an ARR of SEK 63 million, and for those of you who've been following us, for a while, you know that we did have an ARR slightly above that. We don't, in Q in August of, Q3, August of 2023. Now, we moved our ARR measure to being an average of the three months in the quarter, and then times 12. And the reason for that is that in August, we had a spike, which turns out to partially be because of a, a glitch in, in, or, a fraud attempt in the WhatsApp channel. I can say that WhatsApp is back now, we see growth again, up to those levels in, during, Q1.
But in Q4, there was a lot of fixing that channel, so the channel was out for a while. So the WhatsApp channel is partially chat into, so offline chat, you could say, so it's messaging, but it's also the fact that a lot of people in the US who have contacts into Latin America, they use WhatsApp as a communication channel. They just pick up the phone and use WhatsApp. They don't necessarily use traditional phone or mobile phone calls. So SEK 63 million, quite proud of that growth in ARR over the year. And the growth in recurring revenue was 47%.
We still have this patent portfolio, seeing more and more that we can also now show the value of that technology to our customers, like this testing of the accuracy, which touches three of those patents, and 100% of our revenue is recurring software revenue now. So large, large drop, of course, in our revenues, since moving into a new revenue accounting for recurring revenue rather than upfront, taking revenue upfront, but also, of course, from that we took out professional services, and that was then the beginning of 2021. So we are now 100% recurring software, which is, of course, what we wanna grow from and how we can also produce a good enough gross margin going forward, or actually an excellent gross margin going forward. So last slide for me, operational highlights.
We signed a number of annual agreements on SaaS, and on licenses, during the quarter, you see some of them, A1, BankID, Banco BPM, et cetera, on the side here. We also saw then that Telefónica presented an integrated Teneo-powered LLM pilot. This is a pilot that did sentiment detection. So this is one of the first usages of OpenAI, Azure OpenAI, in a enterprise setting. So there's many test cases with this, but in this case, it actually also powered phone calls, real-life phone calls, so quite interesting. And there's evaluation primarily on the cost side on that, which also takes us into a product that we launched during the quarter. We'll come to that at the bottom part of this page.
HRSA, here in the US, launched a live generative AI use case in Teneo, so that's a chat-based, so similar to something like the one that Klarna uses, but of course, managed by Teneo with all the enterprise stuff that we have in the solution with role-based access controls, so not everybody gets to see what's in the RAG. RAG is the database that you use in these cases, and not everybody can now go in and see all the salaries or all the whatever that might be that you put in there, which is the case if you use a standard RAG, for example. So, successfully also launched, we closed it in the beginning of the Q1 2024, and thank you very much, shareholders, for this. It was oversubscribed quite heavily, the rights issue.
Obviously, very happy for the voters' confidence in that rights issue during December, January. Then, Teneo.ai, we also unveiled what we call Adaptive Answers, and this is where essentially you don't get the same answer every time. So we use the LLM systems to provide new answers to customers on questions. So the answers evolve and sound a lot more human-like if you happen to call twice, three times, four times about the same thing, or you have different issues, you might be an IT support engineer. So this just makes it much more personalized customer service in the contact center, also powered by LLM. We also launched then what we call Generative AI Orchestration.
It's built on something called the Stanford FrugalGPT Principles, and this enables a whole 98% cost reduction. So essentially sending a request into. You could take an example from Sweden. There's examples from everywhere. The Municipality of Uddevalla, they opened up a chatbot, and all of a sudden, they were paying SEK 1 million for GPT usage in one month. And of course, that's not something they had in their budget, so they had to shut it down. What this platform does is if you develop that chat on top of Teneo, you will be able to steer which LLM model you send your request to.
So some requests, you can just send them in, not even to LLM, you can use the hybrid LLM model from Microsoft CLU, which is in fact, 10,000 cheaper than sending it to GPT-4.5, or GPT-4 Turbo, which is the one they would use through the API. So very much, a solution to steer the traffic and make sure that you use the most cost-effective, AI system in the background. And we also then unveiled the advanced RAG solution for enterprise AI. We call it Teneo for RAG, and this is, how you can use RAG in a complex customer service operation.
And then, essentially, what it does is it builds an answer platform, but then the answer is actually then coded into Teneo, so in TLML, and it will then be able to respond just as snappy as you found on that call we just did to Microsoft Support here earlier. So typically, on the, you cannot use a RAG solution or an LLM-based solution for picking up the phone because the response times are way too slow, and there's too many components involved, but this way you can do it. So quite an interesting product that we launched, and that's something that we push quite a lot right now in the, in our marketing, as you might see on our webpage. So strong growth in recurring revenue, the full year 2023, it's 47% year-over-year. We're quite happy with that.
Obviously, bigger than the market, but we are growing within our customers, and that's what we're quite happy with. And this year is all about now getting some new logos during 2024 to continue that growth. With that, I'm gonna leave it over to Fredrik to take us through the numbers in more detail.
Thank you, Per. So let's start with some financial highlights. So starting off a bit with for the full year 2023 and Q4 2023, we experienced a solid growth in recurring revenues, as Per already mentioned, and of course, also significantly improved EBITDA. And Per also highlighted the fact that we had 100% recurring revenues in Q4, which also is manifestation that we are now a pure software/SaaS company. So very happy with that. Coming back to recurring revenues, we experienced a growth from SEK 15.7 million in Q4 2023, versus SEK 12 million for the same period last year. So a growth of 31%.
And Per already mentioned for the full year, that number was 47%, so significantly higher than the growth rate in the market. And the growth has been driven by existing accounts, and that's also evident in our strong net revenue retention rate of 145%. In Q4 2023, gross margin also improved. The gross margin also for the full year improved, obviously. But in the fourth quarter, we grew the gross margin from 55%- 77%. And that despite, as you can also see then that we had higher revenues and recurring revenues in Q3 versus Q4, due to the reason that Per mentioned. But volumes are back now again.
And, despite then, you know, a bit softer volumes versus Q3, we also had no commission in Q4, which also then improves our gross margin quite a bit. I think that is also something worth mentioning a bit, kind of on outlook going forward. We also have the same commission plan for the coming year, 2024, as we have had in 2023, for existing accounts. And that basically means that, in order to for a salesperson to get commission, you have to have a growth of roughly 10% per quarter. So adding up 10% every quarter, it's more than 40%.
So that also puts a bit kinda where we, we have our targets, on, on this year, and that's only for, for existing accounts. EBITDA improved to -SEK 5.8 million, and the reason for that is, is primarily higher sales volumes, improved gross margins, obviously, and also that we have reduced our OpEx. So we are, you know, continuously closing the cash cap, as we have also, guided for and, and communicated in our previous calls. Monthly OpEx amounted to SEK 9 million in Q4 2023. We will also look a bit on, how the OpEx has developed, over the period here as well.
Cash position was SEK 15 million, end of year, but, as we will also go through in the coming slide, we will also show that we have had quite a bit of, you know, positive impacts coming in during the first quarter, 2024, which is a bit kind of somewhat improving the cash position. So adjusted, we had a cash position of SEK 55 million, end of year. Next slide, please, Per. Just quickly on net revenue retention rate, which is a new KPI for us that we report in our earnings release and also here in the investor calls.
And it's a bit kind of aligned with, I mean, what Per also mentioned about our strategy during 2023, where we also have reduced the number of staff from 74 employees in Q4 2022. And we are now 61 when we ended Q4. So part of that has also been to grow primarily on existing accounts. And in that regard, the NRR metric is even more important, right? Since we are focusing on existing accounts, and it basically measures how well we manage to grow our existing accounts. So for those that are not, you know, fully on top of these KPIs, you can basically say that a number exceeding 100% means that we are growing in this aspect.
We had an NRR of 145%. I think if we then compare with the B2B enterprise software space, where we think we are placed, it's an amazing number. I think Redeye, who is also on this call, they have a quarterly kind of SaaS universe where they measure this KPI, for instance. I think we ranked number two in Q3 on this metric. I do believe there was a company called Checkin that was the number one, but I think we surpassed them in this quarter, so very happy about that.
And it's an important metric for us since we want to grow with our existing account, and that's a great part of where we also see the growth coming in 2024 and also years ahead, since we have a large potential in our customers. So a very strong development, despite somewhat softer than Q3 API call volumes, basically. Next slide, please, Per. API call volumes on SaaS is a key indicator of how our business is growing. The KPI is an indicator on how our customer application and usage of them are growing. And the more applications slash solutions, you know, covered regions, languages, the higher the API call volumes. So it's a very important key metric for us.
When we know how the development has been on this metric, we know pretty well, you know, how our revenues will pan out as well. And also, as Per mentioned, I will not go into that. I mean, we have changed to an average instead of end-of-period measurements, and that's basically to take away, you know, monthly fluctuations, but also to some extent, you know, seasonality as well. Q3 and in particular, July and August were very, you know, high volume months, and Per mentioned that already the reason for that. But we can also then just say that the Q4 API call volumes were more or less in line with September.
And we also can share with you that we have had a positive growth trajectory on volumes at the beginning of 2024. So we can clearly say that volumes are back. And as Per mentioned, our customer has also, you know, switched on WhatsApp applications again. And we also see lots of positive things, you know, from our existing accounts on new regions being covered, et cetera, et cetera. So a very good and promising start of the year, I would say. And yeah, next slide, Per. Briefly on ARR, and it's also averages that we use here. So that will be the thing we are showcasing going forward only.
The SaaS ARR grew with 24%, year-over-year, Q4 versus Q4 2022, and the total ARR for the same period grew with 31%. So overall, a very strong growth. And obviously, Q3 is a bit kind of distorting, you know, the continuous positive trajectory that we have experienced. But we are very hopeful for Q1 to make this comeback, I would say. Next slide, Per. This slide contains a lot of information, but essentially, it's the recurring revenues split on SaaS and non-SaaS, and also the running numbers on that, LTM-wise, but also on ARR metric.
And we can just say that, as we already said, we had a strong growth in recurring revenues in the quarter, but also, year over year, if we look on all the metrics, LTM, recurring revenues, SaaS ARR, total ARR, et cetera. So, we have already mentioned Q3, which were somewhat distorted. But, overall, I think disregarding, you know, that event, I think it's key, that we continue growing the valuable recurring revenue, and the trend is very encouraging in that respect. And currently, we do see significant further volume potential on our existing customers. And, with those customers we already have, we also see that we can, you know, grow SaaS ARR multifold with many of them. And our, you know, 145% NRR is also a clear signal on.
Our strong organic growth with existing customers. And, we also see continued strong commitment and high interest on our SaaS offering, as Per mentioned as well, from our new customers that we are working on, but also then, you know, our, you know, very happy and loyal existing customers. So, we expect, the recurring revenues to continue growing at a high rate. Per mentioned the Gartner market outlook, and I think, if you then just compare, the 30%, 38% growth rate they have envisioned until 2030, 2031, and then a bit compare also where we have set, you know, the commission plan for this year, it's even higher on existing accounts than what is in the market.
So obviously, we feel quite confident in what we are offering, and that we also have a good situation to continue to grow the coming years with our existing customers, but also on new ones. And next slide, please, Per. OpEx-wise, I mean, we can just say. I mean, the trend is clear here, and we have also, you know, recorded this in our earnings release. We are continuously simplifying our operations in how we work, processes, but also in terms of, you know, simplifying our corporate structures, removing foreign subsidiaries that we don't really need, et cetera. All in order to, of course, reduce cost, but also, you know, improve our efficiency in the way we are working.
On top, we have also reduced the number of employees, which is also then evident in the OpEx run rate. In Q4, we had a monthly OpEx level of approximately SEK 9 million. And we are not, you know, planning any major headcount increases, decreases, you know, for 2024. So that's, I think, the outlook we can give on OpEx. Next slide, please, Per. So this slide is a bit kind of outline our cash position end of year. So basically, we had SEK 50 million in the bank. Following the beginning of the quarter, we have received additional cash. So we have received roughly SEK 5 million from the Spanish tax authorities.
And that's something we in the past have received for certain R&D work that we have done, and then we have got this, you know, paid back from the tax authority in cash. And that we received in Q1. We also, as Per mentioned, closed the rights issue, bringing us in, you know, gross proceeds, so close to SEK 26 million. And on top of that, we have also received roughly SEK 10 million from receivables that we had in the balance sheet as of end of year. So adjusting for this, we basically have SEK 55 million in the bank when we started the year, in the way we look at things, adjusted. So, a good cash position to start the year with, basically. I guess, over to you, Per.
Okay. Thank you, Fredrik. So, we're gonna open it up for Q&A, and I'll stop sharing so that I also can see the screen on my computer if somebody raises their hand. So if you raise your hand, I'll unmute you, and then I see Mark Kroestadt has already raised his hand, so let's see if I can find you on the list. So I need to unmute you first. I will unmute, allow mic, and now you need to unmute yourself as well, Mark, and it should work. Okay, it doesn't work. Let's try again. That is strange. I've enabled you, Mark. I don't know why that doesn't work. Can you try, Fredrik, just clicking on.
Yeah.
On the mic name? Ah, now you're just muted yourself, Mark. Now it should work. So if you hit unmute, Mark. Okay. You need to hit unmute on your side now, Mark, so at the top of your, top of your screen, next to leave or lämna or what it can, might say if it's in Swedish. There's a mic. Okay, let's see if we have any other questions while Mark. While we try to fix this for Mark, but your mic is allowed here, Mark. Any other hands? We have a hand from Forbes. I shall allow your mic.
I have allowed you on mic, and now you can unmute, Forbes.
Yes. Can you hear me?
Yes, good morning.
Perfect. Good morning. Thank you. All right, so my first question then is on the financial targets. You mentioned that, you envision with the sales commission structure that existing customers could grow around 40%, at least, per year. But, could you, could you give some more color on how you envision to reach the SEK 200 million ARR target? Because you will need a lot of, well, new customer growth as well. So please, if you could share your thoughts on that.
Yeah. So we have a few different scenarios for that. Depends, obviously, it could go in many directions, but we have a few different scenarios. And the main scenario is that the majority of that growth is from existing customers. And then we're adding customers during the year to also add to that, but it's primarily the customers that we add in the beginning of the year. And of course, we're hoping, or rather I'm saying that we're gonna close at least one new logo in Q1. And of course, that can then contribute at the end of the year. But the ones that we close at the end of the year are not gonna contribute a whole lot to that. So it's primarily still coming from existing customers, therefore, in our scenario planning.
When you look at growth in existing customers, it might be good to also see that the big trend now is to do what's called conversational commerce. So essentially selling and upselling in the channel and not just providing support or explaining your bill or things that our customers do today. And we could also then do start with outbound phone calls, which we're trialing in a few customers to see where that goes. That's one of the meetings I have later today in Seattle with one of our customers to discuss, is it now from a public perception perspective? Because it's allowed now, there's a legal framework for it in the U.S. Is it allowed from a public perception to be called by the bot?
So that's also another growth area we see in existing customers, which can increase volumes quite a lot. So essentially, two areas, both within the sales side, and then, of course, there's geographies. So moving the solution you just saw from Microsoft, moving that into Europe, for example, which is in the plans for them, that increases volumes obviously quite a lot.
And just to add there, Forbes, I think, I mean, and we also see, you know, tremendous potential in many of the customers that we already have contracted, and we know, you know, what they are, as Pär mentioned, for one of the customers, what they are planning. That is obviously, I mean, the least expensive way to grow. The commission plan is also, you know, set, I mean, 40%. I mean, that's the starting point for the sales guys to get commission, and most sales guys want, wants commission, right? So, I think in that regard, we feel quite confident on our existing customers and what we can do with them.
Good. Good. And to follow up on the financial targets, because you have one, which is to reach the API call run rate of a million by 2024, and that is coming up quite soon, and I mean, it's more than 3 x current levels. But do you need to reach that first in order to reach the ARR target, and what is the sort of connect between the two? Because the ARR target, which is two years away, looks more reasonable to reach, so to say, and the API call target in one-year time is looking quite demanding to reach. So if you could talk about that for a second as well.
I think that you're probably spot on there, Forbes. So what we see is that we managed to reach the other two targets, maybe without hitting the SEK 1 billion. And the thing that has sort of been adjusted here, so we've been adjusting our pricing as well during this, which we weren't quite looking at when we did the goals. So, what we see now is that our prime focus is to hit that one cash flow positive month during this year. That's our prime target. And we see that we can reach that without having that increase in. And there's two factors there. One is we increase the price, the second is that we see that customers use more of the seats that.
We also charge for the seats that you develop, and they use more than we were expecting when we made the model from the beginning. We have more growth there than we were expecting.
Okay. And also, just a quick one, on CapEx, which was SEK 9 million in the quarter and more than twice the rate you have had the previous quarters. Can you sort of explain, where that is coming from and if it's a sustainable level going forward?
Yeah, I think, maybe we should have been more clear on that, but, I mean, now that we are also, you know, in the past, we have been, you know, working a bit in parallel with an on-prem offering and also this new SaaS offering. But, you know, on- from the product side perspective, we are driving this, as a SaaS company going forward, and that also means that we will even more actively, you know, drive customers to, uh, the SaaS platform. And that also means that we have, to some extent, also changed, you know, that we have a higher capitalization on what we actually do on this, you know, cloud SaaS product. And this has been, you know, discussed with our auditors and so on.
So I think going forward, I mean, what you see now in Q4 in terms of capitalization rate, and I think we also mentioned that in the report, that we experience a higher capitalization. I think it's, you know, fair to assume same rate going forward. So essentially, no cash-wise, no impact at all, right, with what we have. But it's more that we put more on the balance sheet, basically.
Okay, thanks. And, final question from me then, a little bit on the market. So, let's see here. Yeah, so, if you could briefly just talk about the main factors which have been holding back your customers in 2023, and on the flip side, the main reason you expect to win new customer contracts in 2024 and the main drivers for those customers to engage with you. Just, some words on that, and I'll step into the queue.
Yeah. So. And that's actually ties into to Mark's first question here as well. So the new customers that we're attracting, what we're positioning now is, of course, showing results, so these recordings, and then we're talking about why we can achieve that, and then we're driving the sales cycle based on that. What usually is so we are tied into an ecosystem of things happening here. So hardly any of our new customers are coming with a on-prem phone system or a contact center system. They're all either they are replacing or they've just replaced that into a cloud-based system, which means that that's a big project from a, that's a big transformation for a company.
It's a big, there's many moving parts when you do that, and we see that and across our customer base as well, that this, this takes a lot of time and effort from customers. So that, it usually controls the timings of these. So we're building the pipeline with, the help of, an external, business development firm, which is building up the pipeline. And then, of course, we're pushing these-- we're juggling these customers to see which ones we can push into each quarter. But we do believe that, with the current pace, we're gonna be able to deliver that probably in the, in the, mid-range, let's say then, Mark. So we've said 4-8 a year. We're we're aiming for at least six, during this year.
So that's the way we would judge the pipeline right now to bring in customers. And maybe if I continue a bit with Mark there, because it's connected. So one of the things is that that's gonna primarily happen through partnerships, Tech Mahindra being quite important in that pipeline, meaning that we have to have a partner with us. The benefit of somebody like Tech Mahindra is, of course, that they can also do BPO. So they can first take over the staff, but also transition the staff that will no longer be needed. So that's quite strong in how they take to market this offer. Another partner is EPAM, which has a lot of a lot of interesting customers in our space.
They're CX-focused, US-based, but also European operation company that recently took over a large, PS contingent, that we happen to know from before. So they, they've been in our market before, actually worked together also at IPsoft Amelia. So EPAM now is putting a lot of effort into this space, coming from having customers in the CX space, so having relationships already with the right people at the customer. So that's why we find that EPAM is an important partner. I don't mention them in the report because really we started that work in, in the beginning of this year. And the Klarna AI statement, I mentioned a bit about this. So, I was hoping to get some attention from this also in the US, because I- it was picked up by Wall- The Wall Street Journal as well.
But I'm not really seeing that here. There's much less focus in. I've only been here, granted, only for one day so far, but met with a lot of customer teams and customer, and it seems to be less focus on removing staff, and more focus on augmenting staff because there's a staff shortage in our customers in the U.S. So it's a bit of a different situation than what we see in Europe and maybe somewhat of a macro difference in the environments. But the Klarna AI statement, we're definitely going to try to use this going forward because our customers do not go out with how many less people are working there.
But obviously, if you automate phone calls, like we listened to the one before here about that Surface computer , then you're not gonna have as many people picking up that phone, and that's gonna be a big saving for you. So, we're gonna be using that in different ways and see if we can spin on that Klarna statement as well. It's important to just differentiate here. Klarna is a fresh technology stack company. They built a very, very good product, which is all, it's a decomposed cloudified product. That's a bit technical, but essentially it means that you can access any point into any point to that product, and that's their main banking system, so way beyond anything that any bank has. Most of them still have mainframe.
But, that also means that there's less integrations, less performance issues, but also they've never offered a phone channel. They've always offered a chat channel. All customers have almost exclusively offered a phone channel before. We have one customer who's going to offer a phone channel now, which is HelloFresh. But, otherwise than that, customers have always had phones. It's difficult to just say, "Now you can't call, now you're just gonna chat." But the chat solution is obviously easier to do because the performance is not as sensitive, i.e., the delay between you requesting and answering, because typically when you chat, you do something else at the same time. So that's the thoughts on the Klarna.
And the fourth question there from from Mark: "In the report, you mentioned several renewals, which of course is a strong sign. Is there price negotiations?" Yes, we are pushing price up. We don't do the upsell necessarily in the in the renewal. The renewal is just the terms and conditions, and then the upsell is really pushing that volume, which is what we measure the sales guys on. We don't measure them on the renewal, we measure them on that volume as it comes. So, we've been pushing price up. We've been pushing up price for components which we weren't necessarily thinking would be a big thing. But for example, we started pricing out, what's called endpoints. So we have a very efficient endpoint where you could deploy a large portion of your volume. You wouldn't really need.
For a million calls, you wouldn't need more than one endpoint. But we saw some customers, for different reasons, having up to 20 endpoints, and now we charge for that. So that's an additional revenue stream, but it's something that we've added into these negotiations in those, in those renewals. Also, we've lifted most customers from the initial SEK 7,500 to the SEK 9,200 and SEK 10,600 a month for the base license. So that depends a bit on which customer you are and what the base license is, but we lifted most of them from the initial price point there.
I mean, important here to stress, I mean, the,. If you, Mark, compare us with many other SaaS companies, which are much more dependent on seats, I mean, our model is, you know, driven more by the volumes from the API calls. And I would say, I mean, that's a key component in that we can actually continue to grow in a relatively easy way as the customer also, you know, wishes. So in that sense, it's also very easy for a customer to continue to grow and add use cases as well.
Thank you, Mark. Sorry that we couldn't join you in here. Not sure what happened there. Are there any other questions? And, in my enthusiasm over doing the business update, we took a bit longer than usual. Nicholas, you have a question. Let me open up the mic for you. Allow mic. Now you can unmute.
You hear me?
We hear you now.
Yes.
Excellent. Well, good morning. Thank you. Actually, not a question, but, and perhaps a little premature to say congratulations, but I would still like to take the liberty of saying a job well done as a minority shareholder, and, impressive determination and perseverance over the past three years. So, just wishing you God speed on this, interesting year ahead, and hope that you'll achieve your objective of cash flow sort of positive a month or even, hopefully, a quarter. So thank you for the job you're doing.
Thank you very much, Nicholas. Very appreciated.
Thank you.
Thank you.
Thank you.
Okay, well, thank you all. I'm gonna hit the sack. It's 2:00 A.M. here in Seattle, so I'm gonna terminate the call now, and thank you all for joining. And again, thank you for the confidence in the company. And thank you, Nicholas, for the positive comments. Thanks, all.
Thank you.
Bye.