Have Keyvan Mohajer, the founder and CEO of SoundHound, for the next 20- 30 minutes at least. I'll get things started, but if there's any questions from the audience, would love to get them. What I'm going to do is I'm going to assume that
SoundHound is new to some folks here, and I want to start from the beginning. You'll see what I mean by beginning, because I mean beginning. I think it's very valuable to understand SoundHound and where it is today in your journey. Hopefully, we'll get to some of the stuff that's more recent as well. I just want to start with, Keyvan, when did you first come up with the idea for SoundHound? Where were you at the time, and how did the story start?
Well, it was more than 20 years ago.
That's right.
Yeah, we were an early pioneer in the field of voice AI. We were working on voice before most of the big tech were even thinking about voice. The history is I started three companies in undergrad when I was 19, 20, and 21, so I had maybe some entrepreneurial itch. I also did really well in school. I ranked first in my class in engineering, and I decided that I wanted to be a high-tech technical founder to a very high-tech company that would make a big impact in the world, and I would spend decades of my life in it. I didn't know what it was, but I thought I need to go to grad school, so I went to Stanford, got my PhD in engineering there.
When I started, I was looking at what would be the next big change in my lifetime that I can work on, I asked myself, what do they have in Star Trek that we don't have? The obvious things are spaceships and replicators, Holodecks, teleportation devices, great products, but maybe a little far-fetched for my lifetime. The one that was less obvious was voice AI. In science fiction, we talk to robots. In real life, we barely had a dictation software that you could use to write emails. I thought voice AI is going to happen for sure. It's going to happen in my lifetime. I wanted to be a part of that transformation. With that thinking, I chose my PhD thesis to be in speech recognition and machine learning.
Studied that for four or five years, then with my co-founders in a dorm room, we started SoundHound to make voice and conversational AI happen in our lifetime.
I want to point out a couple of things, which is Elon Musk right around the same time thought the same thing, he decided to make spaceships.
Yeah.
The other one was that Amplitude is here at the conference, their founder, Spenser, was.
I think he was my student. I was a TA.
Yeah.
A lot of actually founders went through a class that I was TA-ing. One of my co-founders, James, actually was my student also.
Stanford is probably as important to the future of technology as TSMC is. I think we need to put some air defenses around Stanford based on all the stuff that you guys have achieved. You started with that, and then the first couple of markets you targeted were automotive and restaurant. What are your solutions there?
We spent the first 10 years just building the tech, and we wanted to own the whole stack, and we built our own speech recognition engine. We built our own voice AI. In 2015, we unveiled it. It was really well-received. The first industry we went after was automotive. We spent the first four years mostly in automotive, and we did really well. Everyone told us it's hard to sell to car companies, but we did a great job selling to car companies. We want to have dozens of logos that use us, and we are still selling. Q1, our automotive revenue was up 88%. We are doing really well in China. We are not just automotive. We are in TVs, robots, IoT devices, tens of billions of IoT devices that can use voice, and all they need is a microphone and SoundHound.
We think speaking is the most natural way to interact with devices, there will be a day that every device will have a voice interface. In 2019, we went to customer service, we started in restaurants because it was very green, our technology had differentiators. We could handle complex things like food ordering. That was also quite successful. Now we are in double-digit thousands of locations powering, whether it's their phone or drive-through. We did that for a few years. We thought for customer service, it was time to expand to financial institutions, insurance companies, retail, hospitality, and so on. We went there, we identified a path to accelerate that through acquisitions. We bought five companies. Today we are in 25 of the Fortune 100 companies. We are in 12 of the top 15 banks.
We are four of the five airlines, four of the five auto, more than 10 telcos. We are in thousands of businesses, millions of devices. I think we're just getting started.
Yeah. That's exciting. Let's build that up. Let's go through those acquisitions and what they added. What did SYNQ3 add?
SYNQ3 specialized in voice AI for restaurants. They started as a restaurant company that became an AI innovator. We started as an AI company that identified restaurant as a very ripe use case. The combination just made a lot of sense. We bought them a little over two years ago. Great team, great solution, great customer base, lots of experience, they are thriving at SoundHound. They were just ending the COVID era, COVID things went up, they were going down because of the cycles, now they're growing. They're growing at the part of SoundHound. Customers are really happy, they gave us the scale that we needed to make the voice commerce vision come to life.
We added Amelia. What did Amelia add?
Amelia, also AI, they had over two decades of enterprise customer service integration, customers, and data and platform. Together, if you add all of our customers together, all of our acquisitions together, we have more than 120 years of AI customer service for enterprise, customer accumulation, data accumulation, and integration with their infrastructure. That you just cannot replicate overnight. You can make technologies maybe in these days with a few years, but to be deeply rooted in those enterprises, it really takes that long, and we have been able to accumulate that through these acquisitions.
Because they're a horizontal solution, they added verticals for you.
Yes.
Financial services and hospitality, et cetera.
Yes, absolutely. We got a lot of customer acquisition, customer accumulation, but also synergies in products. For example, Amelia had a great developer portal that was a key building block for our OASYS platform that we just announced, our agentic platform. Our last acquisition, last year, was a company called Interactions.
Also amazing set of customers, they had a very unique innovation that is now part of our stack. In enterprise AI, when you automate customer service, you put AI in front of human call centers. Your AI can do a containment. It can do 20%, 30%, 40%, 70%. The more you contain, the better for the customer because they save costs, it doesn't go to the humans. What they invented was they called it human-assisted resolution. Instead of AI failing and passing it to a human, the human is a short-term supervisor for the AI. The caller talks to the AI, and if AI is struggling, it says, "Give me a moment." It goes to a human and say, "What should I do in this case? Quick." The human only spends 10 seconds and chooses an option or resolves the situation.
AI goes back to the caller and continues. We don't hand off the AI to a human. If it's necessary, we take a few seconds of human time instead of five minutes, and that has a huge impact on containment rates and margin for us and for the businesses, for our customers.
Got it. Those are some of the layers you've added to date, I want to get to LivePerson in a second, I do want to circle back to OASYS, I want to ask you a little bit more. This is a good time for me to take the conversation to where we have a lot of conversations with investors, which is the strategy that you have, which is a very unique strategy. I think many of the other companies in this building right now and many of the companies that we cover, when they go do M&A, they are paying very expensive prices for companies that are growing very fast for a variety of reasons, because they want the growth, because they want the team, and that's the strategy that most investors are used to.
Therefore, they have a certain toolkit that they apply to those companies, which includes asking the question, what's your organic growth, right? Oh, you just bought this company, it added a little revenue, but what would your growth have been without that acquisition? I get a lot of questions about SoundHound on that. What I tell people, and I'll just tee you up because I don't want to give the answer. I'll tee you up that your strategy is different. That's not how you pursue M&A. That's not your strategic direction. Maybe you could just step back. We've talked about what these acquisitions have added. You're about to add LivePerson. How is your strategy different than the strategy that most investors are used to analyzing?
Yeah. I think it's wrong to say what's organic, what's not organic, because we are going after the same customers, and in most of the cases, our past acquisitions, they were declining for special circumstances, and we've turned those into growth. I think we should argue that these are organic growth, right? Because we got a fixer-upper that we renovated into an amazing property, and we should get credit for it. Even Q1, we did separate. People were asking, "If you take out all the acquisition, take off all the acquisitions, what was the growth?" It was 88%. Our automotive business unit grew 88% year-over-year in Q1. If you put all the acquisitions together, combined was over 50%. Some quarters were over 100% because of the impact of those acquisitions.
If you look at the pro forma, it's a little lower because of some of them are declining. The way I think about it is that this is a land grab moment, if you want to grab the land, you can go by foot, you can go over the air, and you can go over the sea. If you can do it, you should do all three, right? You don't need to choose. For us, going by foot means direct sales. Going over the sea is channel partners. Going over there is acquisition. We are doing all three because we can. We are doing direct sales and winning logos one by one. We have channel partners that are taking our solution to customers and bringing customers to us, and we are doing acquisitions, and we are winning in the land grab moment.
Yeah. Part of the point is that you don't pay 20x revenue. You've been paying closer to one times revenue for some of these acquisitions. Part of the reason you're doing that is because once you get them, you get to assess which customers you get to keep and which not. You're not trying to reach and pay a lot for a company that has momentum. You're trying to buy customers and, to the point earlier, some incremental technology.
Some incremental technology. We are buying customers and integrations, right? Even if we can win those customers on our own, it might take years for us to integrate with the infrastructure, run a small pilot, and scale. We are accelerating that, and that is worth the price. One more thing that's very important is that in enterprise sales, relationships are very important, right? People buy from their friends. I had this aha moment. I was in a meeting with an enterprise potential customer, and we showed them a demo of our platform. They said, "This is amazing. This is the best thing we've seen." We showed them our pricing, and they said, "This is the best thing we've seen." Every criteria, it was the best.
We said, "Okay, are you going to choose us?" They said, "Oh, we don't want to fire our friends." Right? That was a very informative moment for me, and I said, "You know what? We're going to go buy your friends.
Right.
Right? Acquisition brings relationships that have taken decades to be established, that are really difficult to replicate organically, and it's something that we needed. I don't know if it's something that we need constantly going forward. It's something that we needed in the past, and we achieved it.
Well, in many cases, you get to then back propagate your own technology. You bought the relationship. You can use your technology, which will then delight them, which will then you can retain them.
Yeah. Actually, one thing that most of these customers, these acquisitions had in common was they were using APIs of either the big tech or frontier models, and we were able to replace that with our own models that perform better. Our foundation model beats the big tech and frontier models in all the benchmarks that we've tested, all the customers that we tested, we beat them. We improved the performance, we improved the accuracy, we decreased the costs, and we improved the innovation cycles. There are things that are very unique to some customers that require us to take our engine apart and do innovation within it that you cannot do if you're doing it through an API.
Yeah. The other piece of that that's worth mentioning is on the financial side, on the capital structure side, you get to do a lot of this throughout the market offerings. Most companies at the $2 billion or $3 billion scale do not have the liquidity and the depth of the market that you have. I don't think you've done a secondary issuance as much as you've done at-the-market offerings , which you can do at a very small, if any, discount because your stock is so liquid, which is partially a function, we get back to this later, partially a function that you have a lot of individual investors. Your market is very deep with individual investors, which makes your stock particularly liquid for your size, and that facilitates your ability to buy these companies at this revenue.
Again, it's a very unique strategy. I wanted to delve into that because that's a lot of where I get questions is people just, they haven't seen a company that looks like SoundHound , they're trying to use the wrong tools to analyze it. Which brings us to LivePerson, your biggest acquisition to date. What was your thinking there? How did that fit in that strategy? Let's talk about what they will add when you close the deal, maybe circle back to how we close the deal.
Yeah. We are very proud of all of our acquisitions, they became better, I mean bigger over time. I wouldn't say better because they're all great. They became better. This is the first public company acquisition, which was complicated. First of all, I have watched that company for a long time. I admire them. They were a pioneer in digital web chat. Like if you go to a website, at the bottom right, there's a Chat With Us button. I think they invented that, or pioneered it. Chances are that you have used them when you have used that feature. They have amazing customers, and they have a product called digital, it's a digital CCaaS called Conversational Cloud, that connects the visitors of that website to AI or the human agents of those brands. It's a very complementary product that really fits well with our solutions.
They were in a unique situation. Through a bunch of circumstances, they were declining. It's no secret. It's their public company.
Publicly traded.
Yes. Part of it, when we did the customer health due diligence, part of it was financial instability. Their customers are so big that their procurement assesses the stability of the vendors, at some point they raise a flag that we need to switch. The financial instability caused a domino effect that impacted their innovation pace, the product stability. It caused some decline. SoundHound fixes all of those overnight. Right? The combined company, once we close, will have zero debt, a very strong balance sheet. SoundHound is an AI innovator that those customers really want, and we can fix the stability of their platform. An overnight solution to the main three problems that had caused the churn of their customers. There's one more thing. Their customers want voice. LivePerson only provides digital chat solution.
Their customers ask, "Can you offer us voice?" When LivePerson said no, they had to go and do this multi-vendor strategy, which is not ideal. With SoundHound, we can offer voice to their customers, we can offer digital chat to our customers, the cross-sell opportunity is incredible. Their customers spend over $1 billion a year on digital AI solutions, on AI solutions. LivePerson only gets a couple of 100 million of that. Right? The upside of the existing customer is incredible.
Part of the conversation earlier, I think the quote of the annual report is $50 million, a $200 million run rate, but you're being very conservative on that retention. What you've guided for next year is that they're going to add.
Well, $100 million, at least $100 million of revenue from them that we can grow. We've done customer health due diligence, and $100 million customers are very happy and loyal to them, and we think we can retain at least those. Now, anything that we can achieve on top of that is our achievement because we are going to stop the churn, we are going to do cross-sell and up-sell and turn things around. What we've guided is that next year our revenue should be at least $350 million-$400 million with at least $100 million of growable revenue from LivePerson. The near-term upside is $500 million.
Yep. Yeah. The other side of that is you're going to be able to take out costs. Let's walk through your ability to take costs out from the acquisitions you've already made and potentially from LivePerson.
Yeah. Last year, we identified $20 million of cost saving, which we publicly disclosed in our earnings, because we bought four companies before LivePerson, so there was cost synergies. They have redundant clouds and redundant vendors, and then some redundancy in the headcounts. We have taken those actions. We've identified almost the same amount on top of that. That will take some time to achieve, but it's on track. With LivePerson, it's going to be even more. We don't need multiple cloud vendors and multiple IT solutions and identical systems for HR and so many other things. Some of it is non-regrettable cost saving. Some will take time because there are contracts that need to run their course. We've identified enough to bring us to just the cost saving alone, bring us very close to profitability.
Yep. I'll open it up to questions in a minute. I just want to cover the last topic, which is OASYS. Tell us about what you've built there and why it's important.
Our company. First of all, we are a leading innovator in our space. We were the first company, to my knowledge, that combined speech recognition and natural language understanding in real time, we called it Speech-to-Meaning, back in 2015. Gives us speed and accuracy improvements. Our demos went viral in 2015. You can look on YouTube, "Hound internal demo." It got 2 million views in one day. We are a leading innovator ourselves, but another advantage we have is that we react really well to external disruption in our field, and we reinvent ourselves. We reinvented ourselves when deep learning happened. We reinvented ourselves very fast when Gen AI happened. Then the last 12 months, agentic. Agentic was the last piece of the puzzle that we needed for our business. SoundHound does not need AGI. We don't need to wait for AGI.
We have everything we need. Agentic was the last piece. OASYS is our agentic platform, which will deliver customer service AI solutions to our businesses, to our customers. One of the key features of OASYS is that AI builds AI, and it's self-learning. It used to take us, let's say, nine months to deliver the use cases of a complex insurance company with a team of professional services, engineers, several months and then we go live as a small pilot and scale, and then we have to constantly monitor it to improve it. We can now do that in minutes. Right? What used to take us nine months with lots of smart people and engineers working very hard now can take minutes. We used to go to meetings to win a customer with a PowerPoint and a generic demo.
Now we can go to them with a demo of their own use case. Right? Because it takes us minutes to build what we think they need. With the OASYS, you tell AI what to build, AI builds AI. You can go live, and once people start using it learns from the interactions with the callers and the users and identifies its own weaknesses, proposes fixes and upgrades for a human to approve. We are not there to let it improve by itself. We still have a human review and approve it. Those things used to take a long time. It used to be expensive. It's going to allow us to go to market faster, reduce time to value, improve faster, win more often, and improve our margin because we can just eliminate the professional services.
That's the low margin part of our business that also creates friction in customer adoption. We just eliminate that and we get faster to the high margin subscription revenue.
That makes sense. Any questions for Keyvan? I'll ask you the question that I think you alluded to earlier, but it's one of the questions that we get first, which is, wait a second, Google has voice AI, Amazon has Alexa, Siri, all that stuff. How is it possible that you have a product that's competitive with those guys? Obviously the addition of OpenAI and Anthropic.
Well, we built a voice before, I think all of the companies that you mentioned. We were working on voice before them. It's possible for a small team of really focused, smart people to do better than bigger companies. OpenAI did that with Google. It is possible. We've done that with voice. We have our own foundation model in voice that beats the accuracy and performance of the frontier models and the big tech. That shows up in the public benchmarks that we can measure. It also shows up when our customers benchmark us against those players. The number one reason our customers choose us is the technological superiority. But there's another reason, is our partnership philosophy. Our customers are going through an AI transformation, they need a partner.
They don't want a vendor that says, "Here's an API and some documentation, good luck." We go sit with them. We listen to them. We listen to their challenges. We help them overcome their challenges. We inspire them to dream big, then we help them to achieve those dreams. That is the second reason customers choose us. They can't get that from Frontier Labs or the big tech.
Yeah. Again, you make those benchmarks public, so it's there for everybody to see. Any other questions before I wrap up? Let me just ask you then to put this all together, what is SoundHound today and where are we headed?
We are, I think, one of the very few pure play AI companies that's in the public market. We have a proven track record of constant innovation, and reinventing ourselves, beating the odds. If you look at our history, I think people will write about SoundHound as the one that beat the odds again and again and again. Right now, we are in a special moment because this is our moment. As I mentioned, SoundHound does not need AGI. We don't need to wait for the next breakthrough to happen so that we can go do the next big thing. For what we need, which is AI customer service for businesses and AI assistant for the products, we have everything that we need. Agentic was the last piece of the puzzle, and we have that.
Now is the land grab moment, we are going through air, we're going through sea, and we're going by foot. We are going direct sales, channel partners, and acquisition, and we are doing really well. I expect our revenue will grow at an accelerated pace. I think our profitable measures improve through the cost synergies we identified, and we are getting better at it every day. I think in three years, SoundHound will be a major force in this era of AI, both as a technical innovator, but also in terms of business success.
That's excellent. That's very exciting. Thank you.
Thank you.
Thank you. Thank you, everybody.