SoundHound AI, Inc. (SOUN) Q3 2024 Earnings Report, Transcript and Summary

Hello, and thank you for standing by. I would like to welcome you to the SoundHound Q3 2024 Earnings Conference Call. At this time, all participants are listen only mode. After the speakers' presentation, there will be a question-and-answer session. [Operator Instructions] Please be advised that today's conference is being recorded. I would like to turn the conference over to Scott Smith, Head of Investor Relations. Please go ahead.

Good afternoon and thank you for joining our third quarter 2024 conference call. With me today is our CEO, Keyvan Mohajer; and our CFO, Nitesh Sharan. We will begin with some short remarks before moving to Q&A. We would also like to remind everyone that we will be making forward-looking statements on this call. Actual results could differ materially from those suggested by our forward-looking statements. Please refer to our filings with the SEC for a detailed discussion of the risks and uncertainties that could affect our business and for discussion statements that qualify as forward-looking statements. In addition, we may discuss certain non-GAAP measures. Please refer to today's press release for more detailed financial results and further details on the definitions, limitations and uses of those measures and reconciliations from GAAP to non-GAAP. Also note that the forward-looking statements on this call are based on information available to us as of today's date. We undertake no obligation to update any forward-looking statements except as required by law. Finally, this call is being audio webcast in its entirety on our Investor relations website. An audio replay will be available following today's call. With that, I would like to turn the call over to our CEO, Keyvan Mohajer. Please go ahead, Keyvan.

Thank you, Scott, and thank you to everyone for joining the call today. This quarter, we have seen an acceleration of our mission to voice enable the world with conversational intelligence. We have extended our AI engine customer service offering into hundreds of more brands. We are seeing strong demand across all businesses. We increased revenue by 89% year-over-year and exceeded $25 million in a quarter for the first time ever. We achieved a run rate of over 6 billion queries annualized up over 100% year-over-year and that's excluding the impact of our recent acquisition. We are thrilled to be setting records. While we are seeing massive success in our business, our vision has not wavered. We are pioneering the business evolution of AI founded on our two predictions. First, voice is becoming the preferred way for users to interact with billions of devices and is finally meeting the science fiction quality expectations. Users are adopting it enthusiastically and product creators can embrace it with just a small inexpensive microphone and a partner like SoundHound. Second, AI customer service will become as necessary to all businesses as Wi-Fi and electricity. And thanks to the incredible strides Voice AI has made has made in recent years, we believe users will prefer to have their questions and transactions handled by a consistent, accurate conversational AI agent than by a human. SoundHound is a leader in enterprise conversational AI and we are reaching scale as a pure play AI software company, significantly expanding into new verticals with our AI agent customer service solutions which are embedded deep in hundreds of enterprise brands. Moreover, our swift expansion is into end markets that are projected to grow massively over the coming years. So not only are we growing fast, but the markets we operate in are also growing fast. And just as exciting, our broader addressable market is also expanding rapidly. The business in the third quarter demonstrated our incredible scale. We are building an ecosystem with customers and partners that are increasing their usage of our technology. The breadth of our portfolio is growing and we have seen a massive shift in the way businesses are talking about AI. For most companies, AI adoption is no longer an exploration within their innovation budget, it is a mandate. We are in a position of strength and building a moat with our integrations and arbitration engine across multiple LLMs. And most importantly, our biggest differentiator is our technology which is becoming more penetrated in the market with more businesses, more consumers, and a growing ecosystem of partners. Our AI agents for customer service now span multiple industries and offer solutions that drive efficiencies for companies by improving both employee and customer experience. For employees, we have best-in-class agent led HR and IT solutions. And for companies looking to deliver the very best experience to customers, our AI agents are able to answer queries and initiate all kinds of transactions. This is substantiated by the customers we work with. For example, AeroMexico, American Heritage Credit Union, Aveanna Healthcare, BNP Paribas, Hoffman Financial Group, Nordic Bank, Resorts World Las Vegas, and Sterling National Bank, among hundreds of others. We also work with financial services organizations of all sizes, from regional banks to credit unions to seven out of top 10 financial institutions globally. These customers are using our AI agents for a variety of tasks from flight bookings to bank transfer, to health queries, hotel reservations, and on and on. Our automotive portfolio continues to grow and we are seeing strong traction in the EV space. We won a large deal with a new and up and coming EV OEM in the Middle East this quarter. In total, we have signed four EVs and two of them already have our technology live on the road. We continue to show momentum with Stellantis in various markets, but more importantly, seven of their brands have now rolled out our Generative AI enhanced digital assistant, SoundHound Chat AI with Lancia in Europe being the latest. This technology was the first of its kind solution, leveraging the strength of Generative AI to delight drivers and increase the usage of voice technology within the car. This has been driving strong usage growth over the past two quarters now. Our technology has broken through and we are seeing tremendous interest with other brands that we hope to announce very soon. We've also won new partnerships that increase our presence in the Chinese market in a substantial way. We have partnered with DayinTec, one of the leading Tier 1 software suppliers to automotive manufacturers in China. Additionally, we are beginning another partnership with one of the largest Chinese multinational technology companies in the world, supporting their focus on China's accelerating automotive industry. This is another great accomplishment for our team as we continue to increase our reach globally. And this quarter, we also grew within the Indian market. We signed a new deal with VE Commercial trucks, a joint venture between Volvo and Eicher, to provide them voice AI assistance. Also, we further expanded our partnership with Kia, adding additional Hindi language capabilities to several models. And there is more to come. In a phased manner, we plan to roll out many more languages such as Bangla and Punjabi. Moving on to our AI agent customer service portfolio. I'm very pleased with the way we have grown this business over the course of this year, winning new logos organically, expanding with existing customers, growing our ecosystem of partners and making smart acquisitions. We have now expanded our enterprise customer brands deep into some major industries like financial services and healthcare among others. And we see opportunities in our pipeline for new verticals such as energy and going deeper into retail. Channel and broader technology partnerships are also critical as we build out our ecosystem. We work with Deloitte, Epic, EXL, Fujitsu, General Dynamics Information Technology, Jack Henry, Kyndryl, NICE, Oracle Cerner and ServiceNow, among others, making our world class technology available to a whole host of businesses across a range of verticals. We have two robust offerings for enterprise AI. First is a sophisticated omnichannel conversational AI solution that allows our customers to power seamless phone and chat experiences for their customers. The second is an AI automation offering that helps enterprise IT leaders manage, resolve and document thousands of events in their IT infrastructure to dramatically speed resolution time and lower label costs. It also uses generative AI to build automations and address issues when or before they happen. We often see the opportunity to sell both of these solutions to the same customer and having this comprehensive full package offering makes us the partner of choice for many enterprise brands. I wanted to call out a few critical wins we signed this quarter with our AI agent solutions for enterprises. In telecom, Telefonica renewed for another two years, further expanding our strong relationship with them in South America with opportunities to expand further. In healthcare, MUSC Health was signed to deploy an AI agent powered by SoundHound AI's Amelia patient engagement solution to enhance patients' access. In insurance, together with EXL, a trusted partner to offer solutions in insurance, we secured a deal to work with Transamerica. In banking, Truity Credit Union also chose us, adding to our growing number of credit union customers deploying consumer banking. In retail, Turret signed with us to improve the shopping experience for its popular fashion brand. In government and military, together with General Dynamics Information Technology, we renewed a federal government contract within a branch of the US Military. Finally, one of the largest multinational payment card services companies in the world renewed for another three more years. Those are some great wins with our enterprise AI agent offering, but we are also excited about our opportunity for SMBs with smart answering. This low touch, highly scalable offering is seeing a strong and growing pipeline that includes individual brands with over 1,000 locations. We estimate this massive TAM to be over 30 million businesses just in North America alone. We have solutions to cover the market from SMBs to enterprise across a growing number of industries, so the opportunity in front of us is very exciting. For restaurants, we continue to expand our offering across drive-thru phone ordering and supporting their staff with Employee Assist. Our drive-thru solutions continue to penetrate new brands and expand with existing ones. While our customers realize efficiencies during the ordering process for their customers, what they are also excited about is the upsell per order our solution is creating for them. We power the AI drive-thru for Panda Express in dozens of locations in multiple states and continue to add new locations. We also just went live with Church's Texas Chicken and the feedback from the restaurant staff is incredible. They are delighted with how much faster they're able to accept orders. We continue to expand in more locations with White Castle. Our earliest drive-thru customer is also our largest deployment. Last quarter we implemented innovation related to speeding up the hardware deployments that we believe can help us scale even faster going forward. Because of our tech and ability to navigate different languages and acoustic variations, we already have drive-thru locations in three continents and we're just getting started. It's also our trusted partners that help make this happen. We are deepening our relationship with hardware and platform partners like HME, Olo, Oracle Micros, Samsung, Sapient, Square, Toast and others to build out a scalable package solution that can improve the speed of rollout over time. The demand for our phone ordering continues to accelerate. We just recently signed yet another top 10 QSR brand in pizza and that makes for three of the top pizza brands globally that we now have as customers. Of late, we are also seeing demand from brands to power their mobile apps. This is such a powerful and convenient modality for so many consumers and we are excited to have signed a new enterprise restaurant brand to power their native consumer facing mobile application with voice AI which includes ordering and business query functionality. As off premise ordering continues to grow, we see mobile based voice AI ordering as an emerging channel, further increasing our addressable market. Whether drive-thru, phone ordering, kiosks, mobile apps or even adding efficiencies in the kitchen with Employee Assist, we are seeing tremendous demand from coffee shops to some of the premier QSR burger and chicken restaurants in the world. Some of the wins this quarter include Torchy's Tacos and a large restaurant chain with hundreds of stores that serves home style food and sells unique retail products. We also expanded within existing customers. We had franchise wins with Firehouse Subs, Five Guys, Panda Express and Beef 'O' Brady's as well as one of our larger pizza chains. And with Applebee's, we are now penetrated in two-thirds of their locations, adding to the expansion we see every quarter with Chipotle and Casey's to remain fully implemented in other locations. Our AI customer service product portfolio is resonating with businesses of all sizes. We are the market leader in phone ordering solutions for restaurants. Additionally, we are proud to say that our phone ordering solutions have crossed the milestone of handling 100 million interactions. That's more than 100 million inbound customer calls exclusively with AI. With that, I wanted to share with you some of our trusted brands that have been using our solutions along the way. Chipotle, Casey's, Applebee's, Jersey Mike's, Firehouse Subs, Five Guys, Habit Burger, Noodles & Co, MOD Pizza, California Pizza Kitchen, Corner Bakery, Blaze Pizza, McAlister's Deli, Schlotzsky's, Cafe Zupas and Blake's Lotaburger. Basically, you can go Monday through Sunday using SoundHound Voice AI solutions to get anything from burritos to burgers or pizzas to pies with speed and convenience. We own our tech, have hundreds of patents, have data from real interactions and nearly 20 years of experience. We believe we are winning because of the data science and machine learning behind our solutions. We have built a competitive moat with our proprietary technology that is creating a massive opportunity in customer service. I'm also happy to announce today that we will have a prominent presence at CES 2025 in January. There we will be showcasing several automotive technologies and AI solutions, some for the very first time. This will include our work with NVIDIA to bring voice generative AI to the edge without cloud connectivity. Our live demonstration at CES will show how the solution empowers automakers to enhance in-car experiences by bringing the intelligence of cloud based LLMs directly into vehicles. We will demonstrate how drivers can have meaningful interactions with an in-car assistant using voice generative AI, as well as seamless access to efficient car control capabilities that understand context without the need for cloud connectivity. We believe technology like this will ultimately redefine the way drivers interact with their cars, setting a new standard for the driver experience. We will also demonstrate at CES 2025 what we believe will be the voice commerce ecosystem of the future. We are creating a new category altogether and are excited to share with all of you the dynamic and convenient ways for people to do voice commerce, from ordering food to buying tickets, making appointments, and completing a range of other transactions just by speaking naturally from an ecosystem of devices that we power. More details to come. We look forward to taking part this year at CES 2025. Before closing, I wanted to give an update on our Polaris foundation model, our multimodal, multilingual foundation model that elevates our proprietary automatic speech recognition technology to the next level. Polaris built on two decades of work. It has learned from billions of real conversations and over a million hours of audio in dozens of languages that SoundHound has carefully accumulated over the years. We've been rolling out Polaris in production and the results are exceptional. We are seeing impressive increases in accuracy while also reducing hosting costs, and because we built it ourselves, we can adapt and improve efficiently and with minimal human effort. I'm delighted to say that Polaris is now powering approximately one-third of all AI interactions that SoundHound handles for restaurant customers, and we expect this exceptional SoundHound developed technology will soon be the AI engine to power all our customer experiences. As we continue to develop Polaris, we will also release a number of groundbreaking features that will continue to revolutionize the way our technology understands speech, intent and the full meaning of conversational exchanges. This exciting engineering innovation keeps us at the cutting edge of conversational AI and allows us to deliver consistently superior technology to our customer base. In closing, we have a winning position in all pillars of our business. In automotive, we continue to take market share. We see ourselves ahead of the competition in the adoption of generative AI technologies and AI in the cloud. We are leading in the EV space and with new companies that are choosing their technology partners to bring innovative products to the world. This speaks to our technology and offerings. In customer service, our AI agent offerings are now more than half of our business and growing at a rapid rate. We are working with premier Fortune 500 enterprise brands and adding major system integrator partners to catalyze our business growth. We've had a consistent vision over the years and that vision is now being noticed more than ever before. Companies, partners, stakeholders and most importantly, users of our technology are realizing the amazing experiences we've created. We are grateful to all the collaboration and engagement to all who play a role in our continued joint success, especially our amazing team. With that, I'll now turn the call over to Nitesh to talk about our financial performance, key growth drivers and business outlook.

Thank you, Keyvan, and good afternoon, everyone. Q3 revenue increased 89% year over year. We had over $25 million in revenue this quarter, once again exceeding 50% growth and we are just getting started. Before we dive into the financials for the quarter, let me contextualize how we view the shifting technology landscape and what we are building here. We believe natural language conversations will be the next major interface in how humans interact with technology and generative AI and LLMs are catalyzing that opportunity at an exponential pace. In this new world order, We believe Voice AI is the killer app. This is a generational shift. We are building a foundation for the next five, 10 and 20 years. Our heritage and innovation uniquely positions us to succeed. The basis of our capital activities and investments this year in particular have been to set ourselves up for the shifting reality. As we gear up to finish 2024 and transition into 2025, it is with this lens that we organize and align ourselves. Our financial position remains strong. Throughout 2024, our capital position has been a source of strength and we want to maintain that strength because it affords us the opportunity to grow, invest strategically, extend our reach and deepen our customer footprint. Our cumulative subscriptions and bookings backlog metric is a measure of customer activity and gives value to our current customer contracts. It continues to grow as we further deepen our automotive partnerships and scale the restaurant business. Excluding the Amelia acquisition, it is double what we reported in the prior year period again this quarter and its inclusion substantially increases our enterprise customer base. The Amelia team utilizes their own measure of bookings and we are midstream in aligning all our key metrics. Suffice to say, this expansion into financial services, banking, healthcare, insurance, and more has a meaningful broadening to our customer portfolio and book of business. As one quick measure, if we combined on a like-for-like basis this metric across the full enterprise, we would now be talking about a customer metric in excess of $1 billion and with the duration of approximately six years. That said, when we report our annual results early next year, we will share more on the most appropriate metrics to describe how we are making progress collectively and comprehensively. With that, let me now discuss our third quarter financials in more detail. Q3 revenue was $25.1 million, up 89% year-over-year. We continued to see double-digit automotive unit growth and we saw double-digit unit price expansion in the quarter driven by our Generative AI solutions and overall product expansion. As a reminder, we are anniversarying a period where we had significant point in time revenue from our largest customer within the automotive business. Thanks to our scale and diversification of customers that contribute more SaaS like revenue, we have reduced our reliance on such large point-in-time deals. Within restaurants, we have continued to scale with customers, signed meaningful new logos, and further diversified product offerings. We now count seven of the top 20 quick service restaurants as customers. The quarter also benefited from the acquisition and at a company level, we have a much more diversified customer geographic and industry base. One key measure of that is customer concentration. Last year we reported 72% of our revenue came from one customer and now that number has improved to just 12% this quarter. Expressed another way, last year our top five customers represented over 90% of our business. This year our top five comprises less than a third. This was an area of risk highlighted by investors in the past and so it seemed important to emphasize the tremendous diversification we have experienced over the past 12 months. In Q3, our gross margin was 49%, down year-over-year due to the impact of the business mix from recent acquisitions, which we expect will deliver significant value over time. Adjusted for non-cash amortization of purchase intangibles, our non-GAAP gross margin was 60%. While the acquired businesses weigh on our margins, they provide deep customer relationships and essential enterprise grade data assets. We anticipate margin recovery as we realize synergies from cloud data center migration, tech stack integration and cost rationalization. Additionally, meaningful reductions in inference costs provide a favorable tailwind for our deployments. R&D expenses were $19.5 million in Q3 reflecting a 53% year-over-year increase, primarily driven by our acquisitions. We are committed to delivering innovative products that exceed our customers' expectations and that enhance their experiences. Our focus includes advancing our voice AI engines, integrating leading LLM architectures to elevate our offerings and investing in high impact innovations such as the Polaris initiative which as Keyvan mentioned is an exciting breakthrough not only for SoundHound but for the industry as a whole. With efficient model training as a priority, we are well positioned to pioneer next generation solutions that push boundaries and deliver measurable results. Sales and marketing expenses were $8.4 million in Q3, an increase of 87% year over year. The increase is primarily due to the impact from acquisitions. This quarter, we also continued to invest in go-to-market both across restaurants and automotive in both indirect and direct. Our emphasis continues to be in demand generation, lead generation and improving close rates with speed and efficiency. G&A expenses were $17 million in Q3, an increase of 146% year over year. The increase is primarily due to acquisitions in particular related to transactional costs that are one-time in nature. Note that these are excluded from adjusted EBITDA and non-GAAP earnings. The increase in G&A also continues to reflect growing investments in system and process improvements and enhancements to internal controls ultimately to drive better standardization and cost efficiencies. Non-cash employee stock compensation was $9.1 million and non-cash amortization of intangibles was $5.1 million in Q3. As a result, our operating loss for Q3 was $33.8 million. OI&E was a positive $1.5 million for the quarter. This includes the interest expense of $1.1 million on our debt. We also had a one-time benefit on the tax expense line of approximately $11 million related to the release of valuation allowances related to the Amelia transaction. Note that this one-time benefit is excluded in our reported non-GAAP metrics. Net loss was $21.8 million in the quarter. GAAP net loss per share in Q3 was $0.06. Adjusting for non-cash acquisition-related amortization of purchase intangibles, fair value adjustments, M&A transaction costs, stock based comp and other non-cash items, our non-GAAP EPS loss was $0.04 in the quarter. Adjusted EBITDA was a loss of $15.9 million in Q3. The year-over-year increase was driven primarily by strategic acquisitions and growth investments we have been making in the business. Our cash and equivalents as of September 30th was $136 million and we have $39 million of debt that remains outstanding from our most recent acquisition. Last week, we announced a new at the market equity program. We have utilized ATMs previously to provide capital flexibility and improve our balance sheet. The current outstanding debt has roughly 15% annual cost, so we believe there are economic savings to achieve. We will be thoughtful about when we execute on the program and as I have said before, our capital position is strong and we do not need incremental capital to achieve a breakeven operating profile. With that, let me discuss our financial outlook. We now expect 2024 revenue to be in a range of $82 million to $85 million. And while we will share more details regarding 2025 in our next earnings call, we thought it would be helpful to lay out an early view. We continue to scale our automotive, AI agent, and restaurant businesses. We are integrating the acquisitions and we see significant revenue synergies and we expect to drive meaningful opportunities for upsell and cross sell across the full portfolio. As such, we expect revenue for 2025 to be in a range of $155 million to $175 million. We also expect to drive cost synergies and continue to rationalize our portfolio to capitalize on the most meaningful investment opportunities. One of our biggest priorities is ensuring we continue to operate with agility and urgency. We want to make sure we aren't short changing opportunities either, so we will balance moving fast to capture market opportunities with driving efficiencies to move towards profitability. In accordance, we expect to achieve adjusted EBITDA profitability by the end of 2025. I'd like to close by emphasizing factors that underpin our financials, which may not be obvious on the surface. Over the past year, we have meaningfully diversified our customer base, product set, and industry and geographic coverage. We have solidified the capital base and extended our reach into what is clearly generational and foundational technological shifts that are only just beginning. Another way to view this is that we are steadily derisking our financial profile while concurrently accelerating growth vectors. That certainly doesn't suggest we aren't elevated beta as we are fundamentally a disruptor. It does though portend that we are better situated to drive sustainable alpha for the long term. I have said several times before that our pathway will not always be linear. That is partly by design because we are thinking exponentially. Thank you. And now, we will move to Q&A.

Thank you. At this time, we will conduct the question-and-answer session. [Operator Instructions] Our first question comes from Gil Luria of D.A. Davidson. Your line is now open.

Good afternoon. Three months ago, when you announced the Amelia acquisition, you talked about the fact that you were looking at various parts of the business to decide which ones you're going to incorporate into the future SoundHound and there was a possibility that you may want to discontinue some of those businesses that you had acquired. With the three months’ worth of work and perspective on that, can you give us a sense for how much of the Amelia business you're going to be retaining, how much you may be discontinuing, and then what's the accounting going to be for that?

Sure. Hey, Gil. So to be fair, we're still in the process of that and you're right, we're three months in. In fact, there's a lot that we've been learning and bringing the team together. I think there's a lot we're excited about that we're going to accelerate. There's product cross sell, upsell opportunities we've talked about. There is integration across our technology stacks that we're going to do. And in terms of the composition directly to your question, they have multiple pieces, one of which is this [Technical Difficulty] and this future of agentic sort of space. They have real integrations that we're going to accelerate and that's one we absolutely will keep. High margin, deep integrations, conversations with customers. I think we talked about also last time where one of the big areas they saw a lot of opportunity was voice enablement with our own proprietary tech and sort of displacement of third parties. That is an area absolutely we will double down on and grow and highly excited about that. There is another sort of maybe more employee facing capability. They have a solution called AIOps. That's one where we're probably going to double down and accelerate investment in. And we think that gets us deeper into many of those enterprise brands where we can provide full suite solutions in employee facing conversations, customer facing conversations. The two pieces we talked about previously that we're sort of still admittedly in contemplation around what's the future holds. One is professional services. The second is related to what I'll just call escalation support, pass through support. The professional services is actually strategic and I think the question isn't binary in or out, it's more about where and how. So for example, doing more of these professional services to go deeper into integrations with large enterprises that have hundreds of thousands of interactions -- monetizable interactions, that stuff, we are -- we see a positive return and we'll go faster. What we want to do over time is standardize that so that you're not really every deployment doing more and more professional services that are at a lower margin clip. But that is going to be a journey. And quite openly we probably don't have the full answer for you but over time in the coming quarters, we'll definitely share progress. On the escalation support, that is lower margin today. And the question is we really benefit from the direct customer connection. We benefit certainly, which is probably an understated value proposition of the enterprise grade data assets that come with that and how we can quickly refine and automate our models to improve ultimate delivery for customers. So that we're already starting out of the gate, learning, growing. We believe that business probably will become less, but it will also become more automated. So I don't have a precise number on what the puts and takes are. we gave a bit of a range next year because we probably still need to keep going through this exercise. A lot of it is going to end up being in some sense industry by industry and customer by customer. But I think ultimately if I lift that answer up a bit to something I've said previously, which is, look, we want to be a -- the profile. We are a software business. We think that deploying bits and bytes is much easier. We're certainly in some of our parts of our business. Outside of the Amelia question you asked, but certainly in the restaurant side, where integration -- tight integration with hardware deployments, scales, drive-thru, those are things we want to look at to be faster. But these are the conversations that are happening every day here. And so we're going to actively every quarter try to give you an update. But hopefully, that gives you a little bit of flavor how we're thinking through things.

Yeah. Appreciate it. And then the second question, I'd like to talk about the technology. If you're talking about a foundational model, could you help us understand how you benchmark that model and what the important benchmarks are? How many parameters was it trained on? How big of a cluster of GPUs? What specific benchmarks and what other models are you comparing it to assess how much it can contribute to making your products better?

Yeah. I'll take that one. Thank you for the question. So we've been building models for almost two decades and we have a ton of data. We have data for training, we have data for evaluation, and there are standardized tests out there. But we also have more difficult tests from real life interactions that are actually an asset to give us a better indication of our performance. We have speech recognition benchmarks and then we have natural language understanding benchmarks. Speech recognition, the benchmark is usually word error rate and there's also latency and speed. We beat -- there are not that many companies actually that provide speech recognition. We are one of the few that has that technology in-house and generally, we consistently beat the big tech in accuracy and speed when it comes to speech recognition. And then in terms of natural language understanding, there has been a shift with LLMs. It used to be, you would measure intent recognition and so on. Now we are measuring hallucination because LLMs really always understand, but sometimes they understand incorrectly. And if you look at what OpenAI does, they're one of our vendors, we have our own tech, we use other tech, we arbitrate across LLMs. But what OpenAI does, they put their innovation in front of users in their consumer facing apps and it can work mind blowingly, amazingly, let's say, 70%, 80% of the time. And if it doesn't work, 20%, 30%, that their audience is forgiving because their audience is getting a glimpse of what the future looks like. But our audience is not forgiving. You cannot be 20%, 30% hallucination when you're providing an AI agent for an enterprise. So we spend a lot of time bringing our own IP into the equation and our own innovation to the equation to reduce that hallucination. So our goal is 100% and SoundHound really shines in that area because of the experience and the data that we have. You asked about GPU costs. So again, we've been building models for almost two decades. We have spent a ton collecting data and building models. So if you compare us to a company that's just starting, they have to go and raise hundreds of millions of dollars to get started. That investment is already accumulated for us, but we are increasingly spending more on GPUs to iterate on the models that we have.

Got it. Thank you very much.

Thanks, Gil.

Thank you. One moment for our next question. Our next question comes from the line of Scott Buck of H.C. Wainwright & Co. Your line is now open.

Hi. Good afternoon, guys. Thanks for taking my questions. First of all, I want to say, I appreciate the breakout by kind of industry vertical that you included in the release. But I'm curious as you look towards '25 and the opportunities within each one of those, what your capacity looks like and maybe how you're moving or allocating resources between restaurants and kind of broader retail or auto. Any kind of color you could provide there would be great.

Sure. Thanks, Scott. So we do see a lot of opportunity and we are seeing our long history of building and scaling our relationships with automotive partners is continuing to grow. We've talked about new partners that we've been adding, particularly on the EV side. EVs can scale faster by the way, so that, that gives us confidence. So we want to keep doing that and driving that on the device side. I think where we're really -- if I were to marginally kind of go our -- where we see the outsized pace, we've talked about this, you see this actually in the Misa business is in our pillar two opportunity. And I'll unpack that a bit into the industry dynamics as you -- as you're talking about. But more broadly, just our opportunity to take those AI agents and deploy them across our ecosystem and customer service partners. We absolutely expect that again next year to be outpaced growth. So to double click the restaurant one, it's one that we are very, very excited about. We talked on the prepared remarks here about seven of top 20 QSRs that we're already working with. Many of these when -- we give sort of our indication of scope of opportunity, we talk about the current contracts that we're working with, with sort of a deployment into tens or hundreds of locations. But the reality is across a lot of those QSR fleets, they have thousands of locations. So we see a lot more unlocking and a lot of time in these restaurants there is largely the same menu set. So you get it right, you get to 85 plus order completion rate, you have the structure in place. If it's a drive-thru, you've got the hardware, packaging all combined, you've got a scalable opportunity. And so, I mean, we've been talking about that one now for several quarters. I think the restaurant opportunity continues to be very, very exciting for us. Certainly, now the new industries we talked about, when you get a large enterprise in a large money center bank, for example, or another financial institution or some of those other enterprise spaces, there's just such high volume that we can continue to extend. And again here we're doing both employee facing opportunities as well as customer facing. So those will continue to be big areas. I mentioned briefly in the prior question about a solution we have internally called AIOps. So it's integration with the internal tech stack that is another one that we think we can deepen our roots with customers into IT budgets and that should allow us to scale and penetrate. Over more of -- if I take more of a medium-term answer to your questions, I think just generically the healthcare opportunity is one we're very, very excited about. Today we're integrating with several different both insurance companies but also sort of front end optometries -- optometrists, dentists, medical centers for appointments for understanding what happened in your report, for getting -- there's just a lot more there. It's the largest industry and more and more of it, we see the demand. So quite openly, I'd say those are a couple of them. But, for us, we're a horizontal platform play. We really believe that we're trying to shift to more and more voice first natural conversations, but now we have an omnichannel play. And so where we see consumers that can benefit from this, we're going to deploy that. And some of our solutions like Smart Answering, actually are easily addressable across industries. So we see a lot of opportunity across restaurants, financial services, healthcare. But frankly, it's about deploying the technology horizontally and getting scale across even more.

Great. I appreciate all that detail, Nitesh. My second one on the increased outlook. Is it something you're seeing from -- directly from one of the acquisitions? Is it something you're seeing from the legacy business or is it just a combination of the two that's pushing that higher, both for fourth quarter and '25?

It's a combination of the two, both separately, but also them coming together. We continue to be really excited about just even in answering to your prior question, like the automotive business has a lot more scale opportunity. We're still a smaller share of the total market and we're disrupting the major incumbent player and gaining share all the time. So we think there's a lot more that we can deploy and grow. On the restaurant side I mentioned, we're already with enough partners to really get a ton of organic growth just from those partnerships. But we're adding more and more all the time and scaling units across. And then I think with respect to the acquisitions, the cross-sell, upsell opportunity is tremendous. So we absolutely do see great opportunity across all of those, both organic and across the M&A.

Great. Well, I appreciate the added color, guys. Thank you for the time.

Thanks, Scott.

Thank you. One moment for our next question. Our next question comes from the line of Glenn Mattson of Ladenburg Thalmann. Your line is now open.

Hi. Yeah. Thanks for taking the question. So in relation to that guidance, I'm curious, I think I saw in the filing that there was a -- for the quarter, you'd be like $40 million this quarter if you had Amelia for the full quarter. Correct me if I'm wrong there and then so if you were to annualize that, that's like $160 million next year. Can you just square away perhaps, that's kind of like ballpark to midpoint of your range. So maybe there's some cushion built in there, perhaps there's some seasonality. Can you just help me explain -- help me understand the guidance in relation to that number?

Yeah. I think there's a couple of things there. One is, like, we talked about, I think in Gil's question, there are certain things we're actually looking at in terms of how we're going to pursue that going forward and what mix is. It's not just about growth. It's growth and profitability. I also indicated sort of trying to get towards adjusted EBITDA positive. So we're trying to be mindful both the growth and sort of margin profile. But I don't think that's totally off. I think there is still some seasonality in our business, but as we're moving more towards SaaS, a greater share of our business is SaaS. So I think you'll see that stabilize and hopefully, you won't see it, but we are continuing to grow. And if I maybe unpack the guidance this way, sort of at the lower end of it to the higher end of it, we gave a bit of a range. We want to be prudent in the estimates we're giving out there, just so we continue to give ourselves a bit of the room to breathe to deliver and outpace and make the right investments for the long term. So all those sort of coming into the calibration around the numbers. And I will say, yeah, you're starting to see some disclosures. We actually had disclosure come out recently with respect to the details of the business. But, a lot of our acquisitions and for us to get it at the prices that we feel were attractive, there's a lot of things that we have to work on through integration to right-size the ship to invest in the right place to get go to market recalibrated. And so we're doing a lot of that real time. And that's why looking at the past isn't always the best indicator of the future. Those are the things we're driving with the integration. So hopefully that gives you a little more sense of how we're driving it.

Yeah. That's helpful. Thanks. And then just on the business side, I think you said seven of the top 20 QSRs now. Can you just talk -- in the past you talked about, with these larger deployments sometimes it's a little harder to get whatever if it's all the franchisees on board or something like that, or just to procure the equipment and the time to install everything and everything. Can you just talk about your sense of how that process is going from the wind to the actual deployment?

Yeah. There are a lot of different flavors and a lot of different point of sale systems and integrations that are required. I'll maybe repeat something that Keyvan brought up on the prepared remarks that we're excited about. Yeah. There were -- you're right, there were kind of -- if I think of the franchise locations, there's you got to have a second conversation with franchisees. So we've now activated and are investing in franchise sales and getting the right people because that sales motion is different than new customer acquisition. And so we are investing to make sure with big partners that we have MSAs with that have large footprints of franchises, how do we have those multiple conversations with support from corporate to deploy. So that's something we're definitely scaling in. And then Keyvan mentioned White Castle, one of our earliest partners, and one of the really exciting things this last quarter was how our team got very creative because one of the challenges was the hardware footprint and retrofitting a lot of their locations. And frankly, in some cases, they have -- you got to get a permit in a certain location to retrofit the drive-thru and that could be a multi-month, if not year process. And our team was really creative in coming up with a more agile approach with a post and a visual footprint that we could deploy it quickly, get it activated, quickly get to high order completion rate and let customers start using our technology. So it's those types of innovations that we're doing that we're partnering with many of the hardware vendors that Keyvan noted for both headsets, screens, all the capabilities that are required for drive-thru. And that's the way we're going to calibrate it. So I would say our pace continues. We continue to try to aggressively drive this forward. Relative to last quarter, I think we just continue to learn so much. And I would say the team is frankly going even faster as we learn. And the partnership ecosystem is getting more and more solidified that we feel confident we're going to continue to scale this. And when I get the last thing, you didn't really exactly ask this, but I'll hit it if it's okay. In terms of the interest and momentum, we brought this up last time that, maybe a year ago, we were going out sourcing customers. And now we really -- and I say this with full humility, like, we believe we're absolutely the industry leader driving this. And so more and more restaurant franchises are coming to us and trying to partner with us. And we have, with limited resources, the constraint and challenge of how do we prioritize all of these. And when you move from top 20 conversations to top 10 conversations to top five, we just have to be mindful and edit it and set the right expectation. Our goal here is to exceed expectations and delight our customers and certainly their end consumers. And so that's just a continued calibration that we're constantly trying to battle. But net-net, I would just summarize that by saying, feeling really good about the momentum. We're investing in the right spots. We think this is absolutely on trend where the world is moving next year, the year after, five years down the road. This is where it's going. And we're just super excited about what we're driving here.

That's very helpful, Nitesh. If I could just squeeze in one quick housekeeping on the backlog. I think you said 2x last year, but I don't remember the figure for last year because you had -- you converted to the new backlog number at some point, I think, after that point. Is there a hard dollar you can give us or just -- is that the only…

Yeah. The reason we -- so last year, as reported, it was 300 -- mid-300s. And then you're right. After you normalize it, it's over 400, which I don't think was reported. So the doubling of it, you could just do the rough math. But I think the bigger point I was trying to raise is now we got to conform with what Amelia brings to bear. And they had a very different way of measuring it. So we just -- we'll share more next time. But it's well north of a billion was what I conveyed in the prepared remarks as a combined business. Really nice diversification, I would say, if I were to penetrate -- it's sort of like if I put Amelia to the side, probably the outpaced growth, or not probably, but the outpaced growth continue to be in the restaurant side as we're signing up more customers, but not only that, but activating more locations. But we continue to see strength across multiple parts of that customer base.

Great. Thanks again.

Sure. Thank you.

Thank you. One moment for our next question. Our next question comes from the line of Leo Carpio of Joseph Gunnar. Your line is now open.

Good afternoon gentlemen. First, congratulations on the quarter. I have a couple of questions. The first question is what's your M&A appetite? As you seek to consolidate the voice AI space, any industries or sectors that look appealing and could you provide a little more color on the energy and retail sector comment? It sounds like you're wanting to explore a bit more deeper in there.

Sure, Leo. Thanks for the question. I'll take the first part of that. Yeah. We -- I may have mentioned this last quarter, the quarter before, like I really believe that having a programmatic M&A muscle is imperative. It's especially in fast moving industries and quickly calibrating new horizons, which is what we believe we're in with this generative AI LLM and conversational AI space. And we've seen the benefits of that getting really attractive strategic partnerships; SYNQ3, Allset, Amelia. And we're much stronger as I compare, almost like each earnings call when I compare to what we said last time, we're moving so quickly and expanding and growing the business. This is the right thing to have as a toolkit. That doesn't mean that we are dependent on it. That doesn't mean that we're just actively sourcing anything. But we do have dedicated team and we have a lot of great banking partners who are and we have actually now that we've done a few, we get a lot of companies inbounds that are sending ideas. But we're just judicious, ultimately for us we have built, and Keyvan and team, in particular, built an amazing technological foundation, deep rooted hundreds of patents, deep tech stack innovators. We're constantly pushing the envelope. And the question is how do we leverage that to go faster? And driving and getting deep customer relationships was one of the major theses this year and we were able to do that both in restaurants, but also into other verticals. So going forward we'll be mindful. I think we're at a good spot. There's a lot to digest and there's a lot of work in the integration of the ones we've done. So we also want to be measured and thoughtful about how we're -- what we're asking of the teams. So we're not actively looking, or in dialogue like, but we're not also going to close up shop or put our head in the fan. We're going to be actively looking. And the lens we do is not always ultimately M&A, it's to understand who else is out there. There's new players constantly and who might be good partners, suppliers. There could be a lot of different angles. And so that, that's just all part of the strategic planning that we do. So that's part one. I guess, on part two, because we do have horizontal solutions, we are going after different -- we can hit different industries. I think you mentioned energy and I forget what the other one was, but yeah, there's a lot of pipeline and there's a lot of need for our solutions, particularly the conversational agent side capabilities that we're bringing. So I think that is an area we'll focus on. These are massive TAMs. Again now that we're talking about, if you think industry wise, healthcare, $10 trillion TAM or whatever it is. Financial services also multi-trillion dollar TAM. Energy, multi-trillion dollar. These are spaces that are ripe for automation and something that we will look at. Our edit always will be governed by capacity and resources and prioritizing highest growth and where our capabilities are best suited to fit.

Okay. And then speaking of partnerships, any plans collaborating with the large AI LLM companies like Anthropic, OpenAI or are there existing relationships with them?

Well, we do use -- we publish that we use OpenAI as one of our vendors and we brought ChatGPT into our digital assistant in cars. And we also have disclosed our architecture that you can arbitrate across multiple LLMs. A lot of the queries are handled by our own models, but for certain things, we can go to, let's say, ChatGPT models to answer historical questions. And then if you have questions about real trends and news, we can use models like Perplexity. So we've tried to build our architecture to be kind of LLM agnostic. And as I mentioned, we are also building our own Polaris foundation model.

Okay. And then in the restaurant pillar, what's the point of the conversations? Last quarter you indicated you're getting a lot of inbound calls and given the roster of restaurants that you've been working with, what do you think is the extra that you need to top -- crack into the top five restaurants, the five on (ph) global chains?

Well, we are in the seven of the top 20 and that number has increased a lot quarter over quarter. So, yeah, we are winning new logos. We also believe we are the leader in AI for restaurants. We recently announced that we surpassed processing 100 million inbound calls purely with AI. And another thing that we have disclosed in the past and we also mentioned during our remarks, is that we are going to bring it all together. So we power automotive and TV products and then we power restaurants and not beyond. We are going to bring those customer service solutions into the devices and we are going to demonstrate that at CES. So when you're driving your car, you're already talking to your car that's powered by our AI. So you ask it to car control, navigation, tune into radio, ask for weather, stock prices, sports course. And if you're also powering the drive-thru -- AI drive-thru off the restaurant, why do you actually have to drive to the drive-thru and get in line? You should be able to talk to your car and you talk to the agent of the car, the agent of the car talks to the agent of the restaurant. It's all powered by us. And you can basically place your order in advance before you get there. So we'll demonstrate that at CES and it won't be just limited to ordering food, you can book appointments, buy tickets, do all sorts of voice commerce.

All right. Well, thanks, gentlemen, and congrats again on the quarter.

Thank you.

Thank you.

Thank you. I'm showing no further questions at this time. Thank you for your participation in today's conference. This does conclude the program. You may now disconnect.