Teneo AI AB (publ) (TENEO) Earnings Call Transcript & Summary

Hello, and welcome to Artificial Solutions International AB Q4 2019 Conference Call. [Operator Instructions] Just to remind you, this conference call is being recorded. Today, I'm pleased to present CFO, Chris Bushnell; and CEO, Lawrence Flynn. Please go ahead with your meeting.

Thank you. And first of all, let me welcome everybody to the Q4 2019 call. As is usual with these things, we will be using a presentation that is available on the Investor Relations section of the Artificial Solutions' website. If you want to find that now, if you navigate to artificialsolutions.com, then under the Financial Report section within Investor Relations, there is a tab for presentations. And the top one of those listed is the Q4 2019 presentation. Today, as you just heard, it's myself and Chris Bushnell, and we'll be walking you through this, although many of the things that are contained within were actually trailed previously in the press release that the company issued on the 17th of January. So to let you get to grips with the basics of the companies, we once again, on Slide 3 of the presentation, we've reiterated some of the fundamentals as we may be welcoming new shareholders to the community and new listeners to our first quarterly earnings call as a result of the 100% successfully subscribed 120 million share option that we've just closed. So from the fundamentals point of view, a little bit about Artificial Solutions, and then we will move directly into the news from the company regarding the trading to the end of 2019. So for those of you less familiar, the company is an artificial intelligence software company based in Stockholm operating in the conversational AI area, specifically addressing the needs of the enterprise market. So if you look at Slide 3 of the deck, you'll see that our enterprise customers are indeed blue-chip multinational organizations that deploy our technology in turn to communicate with their own customers. We're overall about 113 staff, approximately 50% of which are technical, working in either the engineering departments, in R&D or in delivering projects for customers. We are multinational, reflecting the needs of our multinational enterprise customers, and we operate through 11 offices around the world, supported ably by a large network of blue-chip partners who help us deliver Teneo-based solutions, Teneo being the AI platform, natural language platform, that the company develops and provides to customers. Our blue-chip customers are able to address the needs of the enterprise market across all of the nations in which they operate and provide us with enormous scalability and reach. So those partnerships, of course, included people like Deloitte and Accenture, but there is some news in the Q4 earnings, which I'll talk about shortly about extensions to our partnership network. As I said, moving on to Slide #4, we do operate in the conversational artificial intelligence marketplace. There is a number of reports that cover off the AI marketplace and specifically subsets of the AI marketplace. And here, we look at markets and markets and Gartner and as examples, who could [ simply realistic ] upon the size of the market. In short, the value of the market is somewhat difficult to quantify. But it's certainly extremely large in terms of the addressable market that the company has as the Teneo platform can be considered to be like a spreadsheet for language rather than a spreadsheet for numbers. So whereas a spreadsheet like Excel may be used in hundreds and thousands of different companies, does it solve different problems that they have ranging from, I don't know, stock accounting to holiday tracking and so on and so forth, the Teneo platform can similarly be used to help large enterprises communicate with their customers, no matter what the application is. So if they want to communicate them to help with customer service or they want to communicate with them to help sell them more things or they want to have a speech-enabled mobile phone application or they want to have ability to interact with them across Amazon Alexa or Google Home or Facebook and so on. It doesn't matter to the Teneo platform what the channel is or what the language is or what the individual use case is, which means that from an addressable market point of view, the company is extremely well positioned to address the full needs of the conversational AI enterprise market. And here, you can see some numbers relating to that in terms of compound aggregate growth in the 30-plus percent per annum and potentially SEK 15.7 billion by 2024. My apologies. So when we turn to the quarter 4 results, here, we've moved on to Slide #5. There are perhaps no surprises here because the company issued a press release on the 17th of January, where we actually trailed our main numbers. So our revenue and EBITDA numbers were published then. And I -- well, I'm pleased to report that the numbers that are published here following the detailed inspection and our audit are 100% accurate and in line with those numbers. So there are no surprises from a metrics point of view for us to announce. So during the quarter, our order intake number was SEK 17 million, which was a little down from last year and then again, in line with the press release that we announced on the 17th of January. This was the result of the delay to an order of some magnitude, which we anticipated to close during the quarter with a major U.S. customer, a major U.S. customer that comes to us through an existing partnership that we have in place. It is unfortunately the case that, that did happen, and it did directly impact our order intake number. Thus, on the 17th of January, we amended our guidance accordingly. We still anticipate that the project and thus, the follow-on order that will materialize, will come to the company during the early part of 2020. And we are working with our partner to secure that transaction. In fact, I'm actually speaking to you in the middle of the night from California, where I've been over working on some of the details around that and other transactions. So regrettably, we ended up with an order intake number which was below our expectations at the time. However, the nature of our revenue types where the vast majority of our revenue is ultimately earned through usage fees will only come once the company has -- once a customer has fully implemented and deployed the solution. And thus, it was not a huge material impact on the 2019 fundamentals in terms of revenue and EBITDA. There was a small impact, but overall, not that material. And we still anticipate to secure the long-term revenues from that contract during the course of 2020 and beyond. As a result of that, we still continue to grow our order backlog, which is the amount of revenue that we will take in future quarters, which we already have fully secured contracts for. So this was actually up 42% on the same period last year, reflecting the fact that although we had 1 event which caused us to produce the press release on the 17th of January, namely that we ordered in closing times, the underlying trend line for the business was, in fact, very, very strong. Revenues, as I said, were affected modestly to the tune of about SEK 0.5 million during the quarter through the missing of the order. But overall, our gross margin continued to trend upwards as it has through the course of 2019, and we are well on track having achieved a 55% gross margin overall. We're well on track to achieve our market guidance of getting that number up to 70% plus during the course of 2020. And that is reflective of the fact that the company is focusing on revenues such as licenses and such as usage fees rather than our more traditional services, which, in this case, are now being delivered by our extensive network of partners and alliances. And thus, we are able to focus on the larger gross margin business, which is reflecting an overall trend in the profitability of the business that we write. At the end of 2019, we've achieved a gross margin of 55% in the quarter. Thus, the EBITDA, which Chris will talk a little bit more about in the quarter was minus SEK 30.6 million, the adjusted EBITDA referring to transaction costs associated with the reverse takeover in March of last year as well. Again, Chris will talk more about those numbers, so I won't dwell on those as we speak now. Also importantly, within the quarter, we continue to show an increased percentage of our revenues, which are delivered through partners. As I've already articulated on this call and it's documented very, very extensively both in the company description and in the prospectus that supported the rights issue, the company's business strategy is to focus on partners who enhance massively the reach of the company well beyond its 110 employees, because our partners have extensive networks of partners, relationships with large enterprises and also to focus on its license and usage revenues because our partners also have the capability to deliver services direct for our customers. So that in the long term, they are an important indicator of the scalability of the business as they can become capable of both engaging with prospects, securing customer contracts and delivering services to enable production artifacts to be realized. And a year ago, 40% of our revenues were driven from partnerships. And at the end of 2019, I'm pleased to report that, that number was up to 54%. That also helps us not just in terms of scaling the top line growth potential of the company but also by maximizing the gross margin number that I talked to you a little bit earlier about. So moving on, therefore, to Slide 6 of the deck. When we look at the highlights for the quarter, perhaps unusually, the main highlights of the quarter was actually something that was closed subsequent to the quarter end. And the reason that, that is the case is because we had produced a press release on the 17th of January, which in many respects, already trailed some of the order intake and EBITDA numbers that I've just talked to you about. The really good news for the company is that the rights issue was 100% fully subscribed. So the company was able to benefit both from existing shareholders following their money, and we're also able to welcome new shareholders to the community. So I'm really delighted that that's in place, and it positions the company extremely well in order to execute on its growth plans and deliver on the rest of the market guidance, which we'll talk about a little bit later in the presentation. So that was concluded very recently. There has been some press releases, regulatory press releases around that. But yes, in all, an excellent outcome for the company, and I'd like to thank both existing shareholders and new ones. As I've already indicated, the order intake, number 1, impacted by the delay to 1 specific transaction. However, we anticipate that, that transaction will close during the early part of 2020 and in actual fact, the long-term impact of the delay in order intake is not that great from a revenue and EBITDA point of view, and it does represent a timing issue rather than a structural one. And that was exactly in line with what we had trailed on the 17th of January. During the quarter, we added 1 new customer to the equation. However, I would like to indicate that this is an extremely important and significant growth from a brand point of view and from a potential future revenues point of view when we added a -- one of the most well-known Scandinavian retailers to the community. They've embarked and are active already in projects to do 5 full-scale pilots, 1 of which is customer-facing. And all of those will be delivered and evaluated before the end of September 2020. So it's a very, very important transaction for us and a very, very nice brand to add to the mix. Also during the quarter, we didn't just benefit from the addition of new customers. But as is always the case, our customers that were already part of the community extended the use of their Teneo platform. Typically, a customer will not just use the Teneo platform for 1 application or 1 language or 1 channel. And they may well start by implementing a solution in the given language on a particular channel such as the website for customer support, for instance. And then they go on to extend the usage of the platform and, as a result of that extension, we gain further revenues, and that exactly happened in Q4 as it has in other quarters previously. A good example of which is that 1 of our major customers in the U.S., a telco, extended one of their solutions to cover their full 11 million customer footprint from where they were before, which, of course, considerably increases the usage dynamic around the transaction and therefore, benefits our company by increased usage. And whilst they did that in quarter 4, the vast majority of that revenue will actually fall in 2020 and beyond. Okay. Also a highlight of the quarter was the fact that we had signed a new partnership in this case with one of the major robotic process automation companies called Blue Prism. For those of you who may be less familiar with this sector, Robotic Process Automation is the sort of cousin to the conversational AI marketplace. It is where companies are deploying technology such as that provided by Blue Prism to automate processes, which has traditionally been done in the back office of organizations by human beings. So Blue Prism has been one of the leaders in that marketplace, has been enjoying an extremely high growth as a marketplace and probably leads the conversational AI market by perhaps 1 or 2 years in terms of customer adoption. So it's a technology that was deployed a little bit earlier than enterprise customer -- enterprise conversational artificial intelligence. Blue Prism, like the other robotic process automation companies, earns most of its revenues through usage fees, the same sort of revenue model that we see at work here in artificial solutions. And as a result of that, the more traffic that they can drive results in robotic processes being kicked off from their existing customers, the more revenue they will make as a business. And thus, it makes very solid sense for a company like Blue Prism to partner with artificial solutions in order to deliver conversational AI applications, which, in turn, will kick off that kind of growth. So we'll introduce those into the customer base where we will help them drive insurance renewals or telco payment cycles and upgrades for contracts and so on, which would normally have been done by human beings but are automated by the combination of a conversational interface put on the front end powered by the Teneo platform and Blue Prism implemented robotic processes that do that, the actual processing of the transactions within their back end system. So that's a very important partnership for us. It's indicative of the importance of conversational AI in the overall enterprise market. And it's a win-win for the 2 companies as the robotic process automation companies like Blue Prism can increase their revenues. And we, of course, gain extra customers and increase traction over time. So we look forward to a very successful partnership there and the RPA, the robotic process automation, sector being important to the company longer term. Finally, before I hand over to Chris on the numbers. We have a very proud record of being an innovator in the conversational AI platform space, and no more so than in our patent portfolio, which is very foundational in its nature. We've had it valued as we've talked to you previously about by external party at Oxford. And when we did so, one of the patents remained pending relating to our Teneo hybrid natural language understanding methods. And I'm very pleased to report that during the quarter, that additional patent went from patent-pending to granted. And it is another example of how the company is maintaining its competitive advantage through technology innovation in the space. And in addition to that, of course, there is an indicator of the value of the underlying intellectual property that the company has developed over the years and months. So that was really important for us. It's not the end of the journey. We continue to develop intellectual property and software that manifests it, and there will be more patent applications going through over the time, but it was an important milestone for us to add another fifth patent to the list of patents that have been granted to the company over the years. So I'll hand over now, let Chris talk you through the key figures slide, Slide #7, of the deck. And then I'll just come back to you to wrap up at the end of the call.

Okay. Thanks very much, Lawrence. Good morning, everybody. Just before I begin, I just -- I feel I ought to clarify one of the statements Lawrence made about the fact, the numbers have been audited. It is true that all of the field audit work has been completed and their auditors have reported to the Board. But the auditor will not actually sign his certificate until we publish our annual report in April. So I don't expect any changes. There's nothing that's predicted to happen, but he hasn't actually put ink on paper yet to confirm his numbers, so just a point of clarification there. On the key figures, as Lawrence has said, these numbers are all very much in line with the press release that we made in January when we highlighted the shortfall in our order intake. In the quarter, our net sales were SEK 11 million compared to SEK 11.5 million in the previous -- in the same quarter last year, that's flat. But the nature of that revenue change such that there was -- we delivered less professional services, the less profitable professional services business and delivered more licenses and usage, that high-margin business. So we see the overall margin improving considerably from 27% last year to 55% this year. And as Lawrence said, our adjust -- that delivered an adjusted EBITDA number of SEK 30.6 million loss against SEK 35.4 million in the previous quarter. The reason we talk about adjusted EBITDA is that we went through the reverse takeover of Indentive in March last year. And therefore, the accounting treatment of professional fees in the reverse takeover mean that they have to be taken to the profit and loss account at the time at which they are incurred. So we effectively eliminate those costs from the EBITDA and report an adjusted EBITDA number that excludes all of those costs. So we have improved slightly since last -- since the same quarter last year. So being Q4, we also talk about the year. So in 2019, our net sales have increased by about SEK 5 million to SEK 49.1 million. Our gross margins overall have increased to 61% from 45%. It depends a bit there on the timing of when the services happened and when the change happened, but we continue to see that move to the more profitable license and usage business. And then the same comments there around adjusted EBITDA, because we had some of the costs of the reverse takeover, who actually were incurred in the fourth quarter of 2018. And the main impact on our cost, we are a people business that most of our costs are to do with hiring, feeding and watering out 113 staff, and we're in a disruptive technology with some of the best people in the world. So we need to make sure that we retain those people. And as the mix changes to the more sales-oriented people who are more expensive and the places where we go, we're hiring people in the U.S. and Singapore, that has put pressure on our personnel costs. And I think those pressures will continue as conversational AI becomes more and more important and the quality of our people is recognized elsewhere. So moving on to Slide 8. We talk about order intake and order in -- order backlog. As Lawrence said, order intake is the contracted value of orders that we've received from our customers, and that did decrease compared to the same quarter last year. But the nature of it continues to be the same way we get. We're getting longer term agreements. So 32% of our order backlog is now expected to be delivered beyond 2020. And as the nature of that changes, the gross margin that's locked up in the backlog is increasing to 81%. Backlog is -- order backlog is the value of orders that we've received but have not yet been taken to revenue. This is where we have contracts for licenses, for professional services and particularly for usage. And we were very strict about how we recognize the revenue, and we take that from backlog and recognize it. And we only count in backlog orders for which we actually have a contractual commitment from the customer. Now when we get to delivering usage, frequently, the customer will not contract for that in advance. And we get to the end of each month or each quarter, and we count up a number of usage that, that customer has and has used them. We raise an invoice and they charge them. So in the month or the quarter in which that happens, it goes into order intake, into backlog and straight into revenue. And what that means is that there's a -- Lawrence talked about the U.S. telecom provider that's just opened up its solution to its 11 million customers. Now that customer doesn't have a contract for usage. So therefore, it's not -- that future revenue stream to come from those customers is not reflected in backlog. So when you're looking at the second chart there, where we look at order backlog of some SEK 35 million at the end of 2018 and SEK 50 million at the end of 2019, there's probably as much again of I call it potential, where we have customers who have solutions, and they are running, and we know they're going to build use solutions, and it's highly unlikely that they will remove them, because they're doing such a good job. So the actual backlog and potential backlog is probably about twice that much, twice as high. So that gives us a much higher level of confidence as we look forward to delivering our revenue numbers for 2020, '21 and beyond. It's a complicated thing I know, and we'll have a look in future quarters how we can demonstrate that better. But the mix of revenue continues to improve. We continue to sell more -- proportionately more license and usage and the more profitable stuff. So -- and the other pillar of our business, as Lawrence has said, is a stronger partner-led business model. You'll see the chart on the right-hand side there that the proportion that we deliver revenue through our indirect and partner channels has grown to 54%, up from 40% the previous year, and I think it was about 15% or 16% in 2017. So there's this gradual progression where we deliver more and more revenue through our partners, and therefore, the partners deliver less profitable services and we get the more profitable license and usage. And -- but that doesn't mean that we leave the partners on their own. The direct sales team continue to support those partners. And I've been involved in meetings this week where we're working very closely between our direct sales teams and our partners to ensure successful projects. So with that, I'll hand back to Lawrence.

Thank you, Chris. So for me, it really just falls to talk about one final slide, which is Slide #10 of the deck. Previous to the rights issue and the prospectus that accompanied it, the company had published 3 sets of market guidance. One was that the order intake number would grow in excess of the market expectation, and as I've covered earlier in the presentations, market expectation is that the market will grow in excess of 30% per annum. And the company is still projecting that as part of its guidance, that is unchanged. Also, as we've talked about a number of times on this call, as the company focuses on more partner-led business, where the partners are delivering the services and thus, Artificial Solutions is able to focus on the most profitable parts of the revenue chain, i.e., the software licenses and the usage fees, we predict that the gross margin numbers will trend above 70% during 2020. And of course, as more and more usage comes on stream, that becomes an easier and easier target for us to achieve. And we've seen a fairly consistent increase in the gross margin, and we predict that, that will continue. Similarly, the company had forecast since its entry on to the first growth -- First North Growth Market in March of last year, that it would turn cash flow positive from operations at the end of 2020, and that guidance remains in place. In fact, it's been massively facilitated, because the 100% successful rights issue has enabled the company to have sufficient working capital in order to support its growth plans through 2020. It's important to note that -- I do apologize. It's important to note that the achievement of revenues in the back end of 2020 does not have to come from new customers. As Chris has illustrated with that example of one of our U.S. customers, suddenly turning on 11 million users to an application, it's in fact, the usage fees that will increase over the year as more and more of our existing customers come on stream that will actually enable the company to get the most profitable revenues and indeed, to see a step change in revenues as we go forward. And in order to articulate that better to the market, the company issued some new guidance during the rights issue that was covered in the prospectus, specifically relating to how usage fees would become an increasingly important component of the overall revenue mix within the company. And we have published guidance that by the end of 2022, usage revenues would, in fact, represent more than 80% of the revenue stream of the company. And that's important because not only are usage fees highly gross margin, but it shows that the growth potential from customers that are already won and the relationships that we have with our partner channel can facilitate a step change in the overall revenues of the company. Because we're obviously not saying that they're going to become 80% of revenue because other revenue streams are going to decrease, in fact, we're saying the opposite because our earlier guidance is that our order intake would go up by more than 30% but such is the anticipated step change in usage over the next 2 to 3 years that we see it becoming a hugely important component of our revenue mix and a hugely profitable one, thereby facilitating such things as turning cash flow positive. And it's also important to remind ourselves that, as Chris did mention, our customers are implementing these solutions because they're getting a return on their investment. It's much cheaper to have automated conversational AI deliver a solution than it is to have people do it. And thus, the usage revenues that we charge are much more cost-effective to customers. So therefore, there is no incentive for our customers to turn the solutions off and replace them with higher cost alternatives. So the very good news for us long-term is that not only will usage fees grow, but they will be sticky. So we don't just have to keep selling in order to get usage fee growth, we just need to deliver and implement solutions. And there have been some press releases around Skoda and Circle K and so on and so forth over recent times, which have articulated exactly that happening in the real world, where our customers are implementing the solutions at scale. So finally, Chris, do you just want to talk through the financial calendar, and then we'll open up for questions. Chris?

Sorry, Lawrence, I've got on to mute. So our next interim report is due in -- for Q1 is due on the 18th of May, which is a day before our AGM on the 19th of May, and we'll be releasing our annual report, I think it's the 27th of April before then.

Good. Good. But if the operator is online, perhaps we can take some questions.

[Operator Instructions] Our first question is from Milan Radia from Hardman & Co.

So it was really around the usage revenue metric, which is clearly a bold pit of guidance and I take on board the comments you made about the existing customers and so on. How does that work in terms of the breakdown of that kind of usage revenue agreement? So how much information do you have today about the take-up patterns across those customer bases and they're rolling out across? And what level of kind of visibility do you have into the kind of progression of that usage revenue over time to get to that sort of 80% metric?

Milan, thanks for the question. Yes. So as Chris said, one of the perhaps frustrating things about our business is that customers are very content to sign, let's say, 3-year license transactions, and they're also very content to sign up implementation projects with us and our partners. But they're very much on the sort of pay-as-you-go model when it comes to usage. And that's because the logic of their acquisition is that the return on investment is based on usage. So in order to achieve an outlook like that, we actually had to go back through the profile of the existing accounts that we have. We actually had to look at detail about the typical engagement that our customers have and how many interactions they have against the rate cards that we have in order to make a projection like that. And as you correctly say, it looked like a bold statement. But I think the reason that we felt we had to give the guidance is that perhaps it hadn't been well understood by the market, but in the natural factor, the growth in usage fees, which is both high profit and can grow exponentially compared to a license transaction. So the short answer to your question is that it's not an absolute sign. How many calls a telco gets into their call center or a tech giant or a fast food ordering business gets at the back end of 2022 is not known to us. And therefore, we've had to make some projections. But what we do know is that the scale of the clients with whom we and our partners are engaging. And the volumes that those customers are having to deal with in terms of customer interactions that are automated by conversational AI is such that it will be a step change in terms of the revenue of the company if we deliver those solutions, and we see no reason why we shouldn't, in conjunction with our partners. And of course, as time goes by, more and more customers are, in fact, stepping up their usage fees. So as Chris indicated, although we have not done that in the learnings over time, we will produce more information for the market over the usage fees and allow you as analysts to track and predict our usage over time. It will be a really big change, because usage fees today are only around about 15%, 1-5 percent of revenues, but they are going to be transformative. And frankly, that's one of the major drivers for the company to drive up those usage fees and to see a long-term predictable cash flow positive business driven as a result of it. So we don't have perfect vision. It's very much like the mobile telephone business, where originally, everybody was on pay-as-you-go contracts. And then luckily, everybody has moved on to preordered contacts. And we anticipate that, that will happen over time. But today, the vast majority of our customers are on the pay as you go. So we have to make some projections as to what the actual usage figures will be over time. So it's not a perfect science, but I hope that has given you some insight into your question.

Lawrence, perhaps -- Milan, perhaps I could add, it's important to remember why usage will grow within an individual customer as well. As you've seen from the list of customers, they are all major multinational organizations. And it's not just the case of -- that we expect the existing use case just to get bigger and better and, therefore, generate more usage. In any customer, and we know of cases of this already, the customer will take the existing solution in 1 or 2 languages. And we know of customers who already have plans to roll out across 30 languages. But they might also want to take on additional use cases. So whilst we might deliver 1 solution to a customer today or 2 or 3, we know of 5, 6, 7, 8 other use cases within that same customer that can be delivered on the same platform that they have plans to roll out over the years. And the customer might also have wanted to communicate with their customers across multiple different channels. So they might today have a web-based chat kind of solution, but they might want to extend that to a phone app to Twitter to Facebook, whatever. So it's not just making the existing solution bigger and better. And of course, we will do that. There are lots of different ways that we see plans within our existing customers of how they will grow licenses and usage revenues over the coming years to facilitate that growth.

That's helpful. I just had one sort of slightly different question which was around the visibility of deals through partners. And Lawrence, you mentioned you're in California working through the details of the deal that is extensively through a partner. So how much -- I don't know whether the complexion of this kind of direct engagement varies across partners depending on how mature your relationships are with them. But how much direct resources are having to be allocated to these larger, more complex deals? And is that something that will sort of also start to put an upward pressure on your personnel costs and demand a greater allocation of resource in that regard?

So as we work more and more with partners, partners become more and more experienced in doing transactions themselves and increase their self-sufficiency. But it is definitely the case that we haven't made any brands about, that our direct sales force helps our partners as we are the center of excellence. It helps our partners close those transactions, and they are complicated. The licenses and associated agreements that we signed in Q4 with the Swedish retailer, in total, they were over 180 pages in the contract. So it's not an unsubstantial challenge to get that sort of complex transaction closed. However, over time, partners become increasingly self-sufficient, and although we have to put some extra yards in at the beginning, they've become more and more capable of handling the whole transaction itself. So we do already benefit from some transactions that have closed without our company having had to invest kind of any labor in getting the transaction done. But that is not true for all partners, especially a new partner like Blue Prism, with whom we may not have done the transaction previously the first few times, yes, obviously, will take more human effort from our side, from our sales machine and so on and so forth. But back to the essence of your question, so what does that mean in terms of personnel costs? I think the bottom line is more or less nothing, because we anticipated that, that would be the requirement. And you've got to remember that if we're putting in somebody into our sales machine who sat behind a partner, that partner may themselves have hundreds, if not thousands, of people in their own sales machine that are capable of making 10-year parts of the agenda with large enterprises. So the partner channel, of course, it has massive advantages. It has massive advantages through scale. But it does require support to get it going, and that's still the phase that the company is in, but that is in line with expectations. And it does have the other problem that we, as a company, are not necessarily having the direct dialogue with the customer. So we are one additional remove from the customer, and that can make things frustrating as we're trying to forecast with 100% accuracy. What offsets that is the fact that the scale that the channel can give you instead of being reliant on perhaps 1, 2 or 3 opportunities, you may end up with 20, 30 or 40 opportunities in your pipeline. And thus, it tends over time to become a much more stable model. So I don't perceive that the company will be having to add personnel costs or support partners that was not already predicted and is not in the most part already on the payroll.

And our next question is from Simon Granath from ABG.

Firstly, you announced a total of 12 new customers during the year. And so far, I presume that they have accounted for a minor share of revenues, but perhaps you could give some color on how much of 2019 revenue that they actually contributed with?

I don't actually know that the number is possible -- I'll hand to you, Chris, but you're right. Yes, you're right. The vast majority of those 12 customers signed 3-year agreements with us. And that means that during the first year of their tenure, and remember that some of those would be as recently signed in December. So during the tenure, the impact on the revenue numbers in 2019 would be very low. Even a customer that signed, so let's say, in March of 2019, they will only have had 9 months of their license fees taken to revenue thus far. And therefore, the subsequent 27 months of their revenue is still to come and sits in that backlog number that Chris has articulated. Further and perhaps more importantly, obviously, if we sign a customer, it takes time to implement a solution. We broke our own world record by implementing the solution in just under 30 days of work with 1 customer after the end of 2019. But more typically, it takes 3 to 6 months for the professional services to be delivered for the security and audit checks to happen for the back end integrations to take place and for the IT -- customer's IT departments to sign off in order to deploy solutions. So even a customer that signed with us, let's say, in March 2019, we would have actually only seen a very small percentage of their potential usage fees hitting our business during the reporting calendar. So I would anticipate that the actual potential -- the actual number from those 12 customers would have been quite low in terms of their impact on license fees and usage fees. Where they probably would have had more impact is if they happen to be a direct customer and they're taking services from us during the course of 2019. So it would be quite a small percentage, but importantly, when we look to 2020 and 2021, they will already have been licensed for those periods, and we would anticipate a big step-up in terms of the value from those customers delivered specifically through usage. So I'm sorry, I can't give you a precise number. We might be able to come back to you with that. But hopefully, that does articulate how -- what the shape of it would be.

I think it was a very good answer. Thank you for that, and thank you for elaborating on that with plenty of points. And as a second question, you -- we have already touched upon that you have a larger telecom customer that has expanded its number of users by a significant amount. And my question is, will you conduct the same pricing in the new deal as in the old one? And also has been this already gone live or when is it expected to do so? And should we expect a gradual uptick in revenues from this customer?

So the various specifics of that transaction was that it was signed in March of 2019. So the -- all of the commercial terms were already agreed. What happened was that the implementation phase went on during that time. And the customer launched the solution and then gradually scaled up the number of customers who had access to the solution. So they basically made sure that the solution was performing excellently before expanding it to their full 11 million customers. So that's what's actually happened during the period. So the deal was done in March. The implementation process happened and then the customer has gained sufficient confidence to fully deploy across all 11 million customers. So that means that the impact on usage revenues was not really reflected during 2019, because it's like all the other customers, it was a pay as you go contract. So they have a rate card but now that they have stepped up the usage, you're absolutely right we anticipate that in 2020, they and of course, our other customers will pay more for usage. And they'll be very happy to do so because in their specific case, it costs them 1/10 of what it costs them to do with people to have the fully automated Teneo powered conversational assistant deliver the service to their customer. So there is not much impact on the numbers that we've shown you, but they and other customers will contribute much better to the usage numbers in 2020 and beyond.