Sinch AB (publ) (SINCH) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the conference call. [Operator Instructions] I must advise you that this conference is being recorded today, on the 19th of March 2020. I would now like to hand the conference over to your first speaker today, Thomas Heath. Please go ahead, sir.

Thank you very much, operator, and welcome, everyone, to this conference call with Sinch and presentation of our acquisition of Chatlayer. My name is Thomas Heath. I'm Chief Strategy Officer and Head of Investor Relations. With me virtually, in this room, are our CEO, Oscar Werner; and our CFO, Roshan Saldanha. And with those opening remarks, I'll leave the floor to Oscar.

Thank you, Thomas. So we are very pleased to announce today that we have acquired Chatlayer, a Belgian NLP AI company, which we think fits -- which we think fit very well to our strategy. So next slide, please. This slide, many of you have seen many times before, so I'll go over it very briefly. We are engaged in customer engagement through mobile technology. We believe that adding the Chatlayer capabilities will enable us to do deeper and more engagements with our enterprises -- or would enable our enterprises to do deeper and more engagements with their consumers. That's, in principle, the core rationale for the deal and I'll explain to you a little bit how that works going forward. You also see quite a bit of the financial figures we've done lately on this slide. I will not dwell on them for this call. But as you know, we have had quite a bit of a very good development over the last quarters. Next slide, please. So this acquisition should be seen in the context of the strategy we have been communicating over the last quarters multiple times. We believe that our market is going from text notifications, where basically the world has noticed that text messages is a great way to reach consumer in a more timely manner than e-mail. So basically, for a portion of the -- when you want to reach a consumer timely and effectively. Text messages are, in some cases, more effective than e-mail because, as you all know, you will read them faster and you will respond to them much faster than an e-mail. Although it has one major limitation, and that is that it's 160 characters. One major shift in the market that goes on right now is that this 160-character limitation is going away. And suddenly, you have richer forms of messaging, WhatsApp, RCS, KakaoTalk, WeChat, what have you, that gives you richer forms of messaging where you can send pictures and videos; you can have action buttons; you can have conversations in messaging, which is what we group in rich media messaging. That is -- this rich media messaging is -- since it removes the 160-character limitation and enables richer form of communications with consumers, it gives enterprises a large set of new opportunities to communicate with the consumers, which is growing the market, which we have seen in many of our market reports. The next phase is not only sending inbound and outbound messaging, but it's actually engaging in conversations with your consumers. If you're an enterprise, if you send out a text message, you may want to get the response from the consumer. So say that you're sending out a message for, "Hey, your flight is canceled." Then your consumer may want to respond to that message in the same channel. They may not want to call you because they may be on an airport and not wanting to engage in their own phone call. The problem that the enterprise has done is -- the problem or opportunity, if you will, that the enterprise then is faced with is, "Well, how do I respond to these incoming messages?" Given that messaging is in a short format -- in a short text format, then it is relatively easy. It's very possible to automate a portion of those messages, to automate a portion of those answers like, "Hey, your flight is canceled." Then the response back would maybe, "Can you find me a hotel?" And you can answer, you can automate a couple of those responses. That's what Chatlayer does. It is kind of adding the automated response to the channels that we have previously delivered to our customers. If the automated answer accounts all the problem, you would be seamlessly transferred into the call center or contact center of the customer. So not to say that you automate everything, you may automate the first part of the flow, the middle part of the flow or the end part of the flow, but that automation obviously increases customer care and reduces the cost for the customer. So that's what this acquisition is about. Next slide, please. We believe that we see our customer categories on a very, very high level in revenue generation, in service enablement or customer care. So revenue generation may be part of our messages add revenue, they reduce churn, et cetera. Service enablement may be you use the CPaaS functionality to deliver a ticket or reduce cost internally and customer care being actually using messaging instead of calls in the customer care center. Like we have said before, we are largest in the revenue generation and the service enablement space, and that's also where Chatlayer fits -- has the biggest fit even though they also have customer care customers. So operator, next slide, please. We have been showing this slide many times before. So we're playing in the connectivity space, which is making sure that a company can make a phone call or send a message. That is actually making sure that you terminate the message to the right handset or that you can make a phone call, say, from a ride-hailing app. When you call the ride-hailing app, we would make sure that, that call happens and we connect you to the drivers. That's the connectivity part. And we are focused on growing that part. It's a very high volume, profitable part of our business, and we're focused on growing that part, both in messaging and voice, but also seeing video coming up -- come up in the future. We're also seeing, as these new channels are coming, more software is needed because the new channels becomes more complex. So our enterprise customers are asking us for more software in order to handle the newer channels. This acquisition is -- fits squarely into the purple top box here. Basically, we have customers, enterprise customers asking us for, "Okay, when I get all these messages in, can you also help me to respond to them and respond to them in an automated way?" And here, we're adding a component, a technical platform component to be able to offer that value-added service to our customers to do so. Charging model-wise, it will be an additional charge to the connectivity. And so if somebody would have a voice call, we would charge for the voice call, as we did before. And if somebody would use messaging, we charge for transmitting the messages as we did before, but we would add an additional charge as a Software-as-a-Service charge on top of that for also providing the service to respond to that goal to our enterprises. So operator, next slide, please. We have also this, on Slide 6, shown many times that we have -- do acquisitions in 2 categories. It's technology and go-to-market and scale and profitability. So scale and profitability being focused on the connectivity side, where technology and go-to-market being focused on adding services into the SaaS layer, and thereby also charging more -- or charging for different types of services to our customers. This acquisition fits very well into the technology and go-to-market. And we're adding a technology and we're adding a go-to-market capability, which Sinch didn't have at a large enough scale before. And we think it's super, super important that we talk both about technology and go-to-market. So here, we get a technology which is out, which is proven, it is with customers. And we add customers and we add a capability to go-to-market here. So very much similar to the acquisition with myElefant. So we can now take this knowledge about this go-to-market, take the salespeople here, have them explain to our salespeople how do you sell this, and then gradually scale this new type of services out to our customers. So operator, next slide, please. So as you have seen before, these are the acquisitions we made. So this acquisition of Chatlayer follows very well on to the acquisition of Sinch and then Vehicle and myElefant and Chatlayer with the technology -- and the latest one now being Chatlayer in the technology and go-to-market piece. And we're just adding various different components here that we feel are strategic to us and that we have not built ourselves. So operator, next slide, please. So the deal rationale is very clear on this one. Chatlayer is a company founded in 2018. It's a cloud-based software platform for chat and voice bots in multiple languages. They deliver AI for Natural Language Understanding, so NLU; or Natural Language Processing, NLP. And they're a team of 17 people in Antwerp, Belgium, so a relatively small company. And the deal rationale is basically, like I said, we add technology and go-to-market to strengthen our product offering and fuel growth in this area. We're increasing the SaaS value-add to our services to add on top of our global connectivity offering. Very interesting specifically with Chatlayer, they have a partner-based go-to-market model and we're a partner-based go-to-market company. So they are selling to large operators and various parts of go-to-market channels and they're selling their web-based software to them, so these partners can then in turn sell to their enterprise customers. And we see significant future in cross-sell and up-sell opportunities with our customer base. Integration. We have an integration plan for doing this company. As always, it takes a little bit of time, but we think it's great to have this know-how and customers internally. Financially, we'll pay an upfront enterprise value of EUR 6.9 million, and the transaction is expected to close early April. The Chatlayer revenues are EUR 0.8 million and gross profit of EUR 0.7 million. So here you can see the gross profit percent of this type of SaaS play, which is obviously much, much higher than the connectivity play because everything here is a software that they provide in the cloud. And the EBITDA of EUR 0.9 million in 2019, so the gross margin being 86% in 2019 since it is a pure SaaS service. And revenues have been growing rapidly to date. So operator, next slide, please. A little bit then on Chatlayer. You see a couple of selected customers. You see Proximus, one of the biggest Belgian operators. You see Foyer and Belfius, so they have quite a few -- and AG Insurance. They have quite a few banking and insurance customers, which match very well to our customer base. And we see use cases here that we can transfer to the other customers in our customer base. On the channel side, they do web, they do WhatsApp, Facebook Messenger, Twitter, phone and a couple of these channels, so it matches very well to our strategy on the different channels they use. And then on the integration partner side, using Proximus and KPMG as integration partners and integrating into companies like Vocalcom and Genesys and Intercom for transferring the conversation to live agents, if the bot doesn't understand what it does. And then having CRM integrations with HubSpot and Salesforce. So operator, next slide, please. To take a concrete example of how this actually looks, this is NLP or AI. So the -- what you can do with normal messaging, if you don't have AI capabilities is what we want to show it on the left-hand phone here. So basically, you would send a message basically saying, "All right, welcome." If somebody contacts, in this case, KD Insurance, KD Insurance would respond with, basically, "All right." So select one of these 4 options: reply 1 for home insurance; reply 2 for auto insurance; reply 3 for life insurance, et cetera. And then the users can do 1, 2, 3. But as in this case, the user didn't really want to talk about this. So people type, "Well, I want to know where the nearest office is." But without AI and NLP, the bot wouldn't understand that. So it would -- you can only do 1, 2, 3 or 0. So the bot would respond, "You've typed an incorrect response, please try again." And then it types 0 to main menu, and then you get, "this is the main menu, please select your options again." And well, you would make a call to the call center, and a call to a call center typically costs like round about $5. Well, if you have AI and NLP, the thing is it would understand the intent of the user. So instead, the flow would look to us as it does on the right-hand side here. So "Hi, I'm KDbot at KD Insurance. Can I help you with your insurance questions?" The person would type, what is the nearest office? And the logic here is the AI would understand the intent of a free text question from the user, so the bot will maybe answer, "What city are you in?" if it doesn't have location-based services connected. And then the person would -- they would respond, "I'm in Liverpool." The bot would understand that and then respond with, "All right, I have 2 offices, and these are the locations." So the thing here is understanding the intent of the user when it sends in a free text form instead of being -- instead of having to be forced to use scripted flow, like on the left-hand side. All right, operator, next slide, please. This is actually a screenshot from the web user interface. This is a company who have both AI and they have a graphical user interface and an online sign up for customers, so selling AI with an online sign up. So you can actually go to Chatlayer, sign up online, and you can create your bots in a graphical user interface on their web page. This is showing a portion of their web page which is used for training bots. So a customer would see this type of flow. So for example, they would start with, all right, the session start, that is, a conversation start. And in this case, 17 -- 41% of the user -- the first question to this bot was to book a train ticket, 11.5% was that they lost an item, 7.4% was that the first question was departure times and 17% was the bot understood the first -- that the user wanted our contact and 2.1% was not understood. All right. And then after, the second question here was if you booked a train ticket, the second thing the user wanted to ask was about arrival time for 12.5% for the user and 17.6% was arrival location. So this is how the bot understands the intent of the user. And as you can see, it's down at the 2.1% or 7% that the bot does not understand. And then you take the user through this flow. And the good thing here with an open text format is that users can, in their normal conversation, jump from one area to another in the conversation without telling the bot and the bot will still understand you. So this is how, I mean, a web user interface to the customer would look and they can use this interface to train their bots. All right. Operator, next slide, please. So a use case take -- doing AI for Belfius, which is a banking and finance customers in Belgium, one of the Tier 1 banks. It's a mobile app. It's one of the most used banking apps in Belgium. And Chatlayer has helped them with 2 different bots. One bot is answering -- being expert on questions on fees and how to transfer money abroad. So if you want to transfer money abroad, you would get a set of the questions answered automatically. And the other bot, myBo, on the bottom part of this grid is a bot that is trained on insurance claims. So it's basically guiding the user step by step to make sure all the relevant user information is collected on the insurance claim and then possibly being connected to a live agent when you have collected the information basically. So -- and the services rated on a 4 to 5 -- 4.5 case by -- 4.5 rate out of 5 by the end users. So I think it's very, very concrete on the value to an enterprise. Our financial targets, this being the last slide. We have a target to grow adjusted EBITDA per share with 20% per year. And we have a net debt to -- over EBITDA to -- at a target of 2.5. And we did grow adjusted EBITDA per share with 54% in Q4, measured on a rolling 12-month basis. And the net debt-to-EBITDA is 1.7, measured on a rolling 12-month basis. So this is obviously, being a very small acquisition for us, it's well in line with our financial targets. And I think you -- the people who have followed us can see that it's very well aligned with our strategy as well. Obviously, a risk, all right? So we have had a lot of considerations on the timing, especially in the situation we are with the corona and COVID-19. And we believe that this is very well aligned with our strategy and well in line with our financial capabilities even in times like this. But we obviously both with the Board and with our management team have gone through those questions in detail and thought through it in very much detail, and we believe it's very much in line with our financial capabilities. And therefore, we think it's good for our shareholders to proceed with this deal despite the COVID-19 crisis. Okay. With that said, I will finish the presentation and open up for questions.

[Operator Instructions] The first question comes from the line of Daniel Djurberg.

Congratulations on the acquisition. I have a couple of questions, if I may. Just starting with Chatlayer, and the question would be, I think you mentioned like 6 customers, some -- at least 2 very big here, Proximus, Belfius and so on. Still only EUR 0.8 million in revenues for full year 2019, if I read correctly, which to me is a small amount even on a recurring basis. Can you comment on -- is this more like pilots? So why is the revenue as limited?

Very good question. It is not pilots. It is live bots out with customers. Obviously, there are a set of pilots. But you're correct that the revenue is limited. But the reason is that they have been starting with well-defined use cases in smaller areas because that is the way you build bots. You don't try to boil the ocean with bots, right? You try to take a very well-specified use cases, train it to be very good at that, and then you expand gradually from that. So that's the reason. It's been proven over and over time, if you try to boil the ocean in helping to make a bot that answers everything, you will fail or the success rate of the bot will be much lower. So that's the reason. We see great opportunities of growing with these customers and new customers in many different areas, basically. We think it's a low amount of revenue, but we think it's an acquisition where we pick up great competence and great traction to a good price for what we're getting. Obviously, as you all know, acquiring AI companies that are significantly later at the stage will become much more expensive. And therefore, we think for our fit, for our needs and for Chatlayer needs, we think this is a very, very good combination because they need us to be able to scale, and we need them to have the competence. We think the timing here is actually very good, even though I agree with your comment.

Okay. And if I may, just a follow-up. Would you say that this is also important from checking the box perspective in terms of procurement processes for your large enterprise customers that you don't need to rely fully on third-party vendors, huge vendors like offering similar functionalities?

Yes. That's not the main reason to why we're doing this type of deal. We don't make acquisitions for checking the box reasons. Of course, it may come up in some area. Of course, it may bring us then further in that discussion, but it's not the main rationale. The main rationale is we fundamentally believe in that this is a service that is and will be requested by our main enterprises, and we want to gradually build on that. Now we should also say, it is relatively early in this stage. It's not like every customer is doing it. And that's why we're also looking at a smaller acquisition in order to be able to build gradually our knowledge and compromise on revenue in this area.

Perfect. And if I may, being the first in line here, obviously, we'll get questions on the resilience to COVID-19 and corona impact on your different business segments. Perhaps you could comment a bit. Personally, I guess, traveling and ride-hailing will be tough for a while, while e-commerce payments and shipping notifications, et cetera, will be quite positive. So it would be great to hear your view on this so far.

Yes. The first thing is we're taking COVID-19 very seriously. I mean the first -- our first concern is the security and -- of our employees. So we've taken various measures. We're all working from home as of today. But then, it's obviously to our existing customers and ensuring business continuity. And we're putting -- we have put them very gradually, and we have been putting those plans into place, and we think we are in a good shape, and we're continuously improving our plans. And we have set up governance of twice-a-week management meetings where we're handling the -- any aspects of the COVID-19, and it's a wide and deep and serious program that we're doing. We're obviously also being in this type of times, we're more -- we're considering options on OpEx and being more conservative in many different areas in order to make sure that we can see what's going on here in the market. We're obviously looking at monitoring our traffic. Again, we're not giving any forward-looking statements, but we're looking at various different segments, like you say, obviously, travel. Travel is -- there are a couple of segments that are tougher, like travel is tougher. On the other hand, there are other segments which are better, which can be the collaboration tools or online sign up to the web made and things like that. It's obviously stronger in these times. Interestingly enough, with Sinch, we're a -- with 40 billion transactions and global, we're at a pretty good cross-section of different industries. And you see trends in various industries that are some are tougher and some are better. But we're obviously evaluating our options and looking closely at traffic and having good plans and contingency plans for these situations.

Hopefully, we don't need to talk so much about corona in the future, but who knows.

Thank you.

The next question comes from the line of Predrag Savinovic.

A couple of questions for me as well. First one, out of your existing client base, could you elaborate a bit on how large part of your existing clients you think could benefit from the Chatlayer service? As in percentage, 40%, 30%, 20%, any idea?

Well, it's a -- the question depends on the time span of your answer.

Say 5 years?

I think -- yes, if I answer in general, if I think if you're doing mobile communication, a large portion of the market is moving to handle not only notifications but also conversations in a message-based format with your customers. If you move into conversation, it is logical to assume that a large portion of these customers will want or will evaluate to have a portion of those conversations, especially in the more repetitive tasks, automated. So that would indicate that in the long run, a large portion of our customer base would be interested in these type of services. If that is 30% or 40% or 50% or 60%, it's very hard to tell, but it's not in the 10% range, I would say. It would be my assumption as of today. Now enterprises, this will take a long time. Enterprises take quite a bit of time to change their communications to their customers. So I think you're going to see -- we were a text messaging. The first way we're going to see a gradual shift on to newer and richer messaging formats, and the second way we're going to see is a gradual shift to more conversation-based communication. Yes, I hope that answers your questions in a general way. It's hard to give a very specific estimate there, but I think you understand the logic here on what we think is happening in the market.

Yes. For sure. And looking at LivePerson then, for example, which we regard as a peer in this conversation vertical. And we have seen some signs of a toughening market for them, lower ASPs, according to various reports, not to themselves. But do you share this view? Or is it rather that actors like Sinch, which can bundle multiple services together, are the disruptor in this market? How are your thoughts there?

So LivePerson, it's a good reference point. I think we're a little bit in a different market and approaching it a little bit differently. We have a large messaging-based business. So we have a vastly higher amount of volumes in our messages, in our outbound messages than LivePerson. And they have a much larger conversational capabilities, and they come from the web chat space, right? So they come from web chat into messaging, and we come from outbound messaging into conversations, if you will. I don't want to comment specifically on their performance and things like that. But I think this acquisition to us is in a low-risk way adding conversational capabilities, which I think is good. Their business model is more to be 100% in it. And therefore, I think you should see the risk levels a little bit different. It is also so that they are focusing more on the customer care side, where you have to replace the care center, which is a business model which takes a little bit longer because replacing a customer care software of an enterprise takes a long time. We are focusing more on smaller integrations in the service delivery and marketing or revenue generation side, where it's more value additive, and you don't have to replace the customer care center. And our strategy is to partner with the care centers. We're the bots providers. We're the automating providers. And then we hand over the messages to the care center when they are. So we have a little bit of a different tactic of approaching the market. But again, I think adding this capability is good, and especially since the financial implication to us is very limited.

All right. Super. That also answered my follow-up questions on whether you would move to this adjacent technology, which could be then to handle the contact center themselves. But another question then on customer care and marketing, that is generally a good fit for conversational and in addition to other use cases that you also mentioned. Do you see any traffic at risks from consumers being more hesitant on how data is handled and data privacy, for example? Or do you see these type of services more being that customers themselves opt in for the engagement?

Specifically relating to Chatlayer, you mean? Or is it a general question?

To Chatlayer and in general, both, if you may.

I mean data privacy is what's most important to us where now, we want to protect consumers and consumers' data. We welcome data privacy regulations because we think it's ultimately good for consumers. That's our basic standpoint. And we implement those processes in all our systems and processes and organizations. So that's our basic assumption. Then -- so -- and then our market is driven, but customers still want to interact with their brands, right? Because that's a need they have. If you bought somebody -- if you want an insurance from somebody, you need to interact with the brand. And as long as the brand and all its suppliers is keeping to the data privacy regulations, I think that's good, and that's exactly what we do. Of course, that limits people like us in certain areas from doing certain things, which we think is good because that protects the consumer. So I see that as being a good thing for the market and then in the long run, a good thing for all the players because we're complying to the rules and to the policies that our consumers want. And specifically, when related to Chatlayer, same thing there. Well, we are complying to the rules and regulations that exists, that limits a couple of use cases, but that's a good thing because that drives a more healthy market in the long run, and that's the type of markets we want to be in. It's kind of markets that grow with the long term with a good confidence of consumers.

The next question comes from the line of Ramil Koria.

So a lot of good questions have been asked. Let me do the nitty-gritty housekeeping questions here. Just first off, how is this actually sold? Is it sold per seat? Or is it per sort of -- is it based on throughput, termination fees or what have you?

The charging model, you mean, the monetization model?

Yes, for Chatlayer, specifically. Yes.

Yes. So Chatlayer specifically is based on monthly SaaS fees. So you're basically signing up to a certain capacity and you pay a certain amount of money depending on your capacity and the number of bots and the number -- the amount of technology that you have. So just think about it as a pure-play SaaS. Then Chatlayer today does not monetize the traffic because that's done by other partners. Coming in with us, we would obviously send the traffic via our channels, so we would add a little bit of monetization just for the voice call or for the message transaction because we would send it via our channel. So that's a small addition. I don't think you should overestimate that as of today. But Chatlayer, as such, is a pure SaaS fee. One can imagine that you would, in the future, go to a more success-based rate. So the easiest example to understand is maybe customer care, where a call to a customer care agent costs $5, and there are people in the industry charging between $0.5 and $1.5 per automated customer care contact. So you can move to a transaction-based model, but that is not the revenues of Chatlayer as of today.

And just -- I'd love to hear a bit on how you source this. Have you come across Chatlayer in your sort of day-to-day operations? Was there any competition for the acquisition, et cetera?

So I do not comment on competition for acquisitions, but I can make comment on how we source it. The sourcing is we identified this in the business plan, last year's business plan, but also in this one. It was a thing we wanted to add to our offering. Then we had our M&A integration team sourcing and speaking to a set of different companies and interviewing and talking to a set of different companies and trying out a set of different companies. And of these, Chatlayer stood out, partly because it was the right size level, partly because it had a very automated flow, so -- and partly because it was AI that was productized. A lot of the companies in this space are more services-oriented. But we, obviously, being a platform company and platform player, we want to acquire companies in our platform place as well. So that we think, all right, Chatlayer was the best company that we saw during these periods. In this space, there are a lot of different companies. We have obviously not met every single company on the globe, but we think this really stood out in many different ways, and that's why we proceeded.

Yes. And then just one final question, unless you had something to add on the former one, Oscar. But just you mentioned that you have a business plan, which you're sort of following through in terms of what you want to add to your offering, et cetera. Could there be something more shorter term on the technology side? I mean we're seeing one of your peers internally developing an IoT solution, which I know Sinch previously has been working on. One of your peers have added payment solutions, et cetera, now that the two-way communications is emerging and sort of conversational commerce is becoming increasingly, call it, interesting for players like you guys. Could there be that you add new technologies? Or is focus from here throughout 2020, say, to sort of develop the current product portfolio?

I mean we approach the market both from an M&A front and from an internal development front. We always make a buy analysis when we approach new segments or a new product area. We have an active M&A agenda, and we have an active internal development agenda. So -- and our M&A agenda is, like we said before, in technology and go-to-market and scale and profitability. So yes, there are multiple opportunities to both do more connectivity acquisitions or technology and go-to-market acquisitions in a market that is as dynamic as ours. For each of these, we're obviously considering, do we develop internally or do we do it via acquisitions? So yes, there are opportunities, but we also consider our internal development. And I can obviously not comment on the specific areas as you understand, but it's a very dynamic market. It's a very interesting market, it's a very fast-moving market and it's a market where you can make a lot of money today. So there you have to be careful -- we're considering our options carefully, and we're executing gradually based on the opportunity we see.

[Operator Instructions] Dear speakers, there are no questions over the phone.

In that case, thank you very much from us. And thanks for taking the time and listening in, both those participating on the phone and everyone that joined us through the webcast. Oscar, any final remarks?

No. I want to thank you for your interest in Sinch. So thanks a lot for participating in this call, and thanks a lot for asking all the questions. We love the interest, so keep on chasing us with good questions. Thank you.

That does conclude our conference for today. Thank you for participating. You may all disconnect. Have a nice day. Dear speakers, please stand by.