CrowdStrike partners with Salesforce to secure AI agents and applications
On Thursday, 11 September 2025, SoundHound AI (NASDAQ:SOUN) presented at the Piper Sandler 4th Annual Growth Frontiers Conference. The company’s CFO, Nitesh Sharan, outlined a forward-looking strategy focused on technological innovation and strategic acquisitions, while also addressing challenges and opportunities in the rapidly evolving conversational AI market.
Key Takeaways
- SoundHound AI’s Polaris model shows significant improvements over competitors in accuracy, speed, and cost.
- The acquisition of Interactions aims to leverage its customer base and technology for accelerated growth.
- The company is targeting strong double-digit growth by expanding in existing markets and entering new ones.
- Strategic partnerships are key to deploying solutions in sectors like automotive and restaurants.
- SoundHound AI has a bookings backlog exceeding $1 billion, indicating substantial future revenue potential.
Financial Performance and Future Outlook
SoundHound AI is focused on maintaining strong double-digit growth in the long term. The company has shifted towards a usage and consumption-based revenue model, which aligns growth with increasing customer interactions and expanding use cases. With a bookings backlog of over $1 billion, the company is well-positioned for future revenue generation. The second half of the year will focus on achieving set targets to meet financial expectations.
Operational Updates
The acquisition of Interactions for $60 million, with additional earnout potential, is a strategic move to enhance SoundHound AI’s enterprise offerings. The integration of Interactions’ customer base and technology is expected to create synergistic growth opportunities. Additionally, SoundHound AI is reinvesting in its Amelia product line to drive significant deals and partnerships.
Technological Advancements
SoundHound AI’s Polaris foundation model is a cornerstone of its technological strategy, offering 30-40% better accuracy compared to OpenAI’s Whisper and Google. The model’s deployment across restaurant solutions demonstrates its adaptability and effectiveness. The company’s proprietary algorithms and access to production-ready data are key differentiators in the market.
Market Opportunities
In the automotive sector, SoundHound AI is expanding into electric vehicles, focusing on increasing the average selling price per unit with new product sets. In the restaurant sector, the deployment of Smart Answering and drive-thru solutions continues to grow, with significant potential for further expansion. The company is also exploring voice commerce as a new initiative to enhance consumer experiences across sectors.
Q&A Highlights
During the Q&A session, SoundHound AI emphasized the importance of channel partnerships with companies like Square, Toast, and Oracle for restaurant deployments. The company sees its unique technological moat as superior to both legacy providers and big tech competitors. Customers are attracted by the potential for cost savings and revenue uplift through AI-driven solutions.
In conclusion, for those interested in a deeper dive into SoundHound AI’s strategic plans and market positioning, the full transcript of the conference call provides comprehensive insights.
Full transcript - Piper Sandler 4th Annual Growth Frontiers Conference:
Tim Fish, Analyst, Piper Sandler Research: Okay. All right. Good afternoon, everyone. Tim Fish with Piper Sandler Research. Thanks for joining us here in one of the last sessions of the Tech Conference for this year in Nashville. We have SoundHound AI here with us. Thanks for joining us, Nitesh.
Nitesh Sharan, CFO, SoundHound AI: Thanks for having me.
Tim Fish, Analyst, Piper Sandler Research: We recently picked up coverage of this, for those unaware, but I wanted to chat about Polaris first, just to get right into it. How long is it going to take to move all of your solutions on top of Polaris? What makes Polaris different than what’s out there?
Nitesh Sharan, CFO, SoundHound AI: Yeah, maybe I’ll go backwards on those questions and first start, like, our differentiation in general, our history has been around technological differentiation and breakthrough innovation, and Polaris is just this latest horizon of that. Polaris is our own foundation model. We really have differentiated ourselves on all parts of the voice AI conversational stack, from speech recognition to natural language understanding, generation, and text-to-speech. This really plays in the speech recognition side where when we benchmark it against the likes of OpenAI’s Whisper or Google, we see 40%, 30%, 30%, 40% better accuracy. We do it with better speed, and we do it at a much smaller cost footprint, which is really important, especially as you deploy these into production use cases where sometimes you have a smaller hardware footprint to utilize the technology. It is now deployed across all of our restaurants.
When we adopted the original platforms with our restaurant, we went on a journey to migrate, and over the course of just several quarters, we were able to migrate all the restaurants. Now, within the enterprise space, we’re on that journey also. While every environment’s a little bit different, I think thinking maybe quarters, not years, is probably the horizon of how we’re trying to do this. We want to make sure that we’re using the best services and best solutions. Lastly, I’ll just say, our differentiation on Polaris, but I think I’ll represent this story more broadly to how we innovate across the whole tech stack, is that we have unique things about SoundHound AI that I think other players don’t have. Number one, we always look at it from a unique architectural place. We have architectural differences.
You sort of find the best of where machine learning makes sense or software development algorithms make sense. We have our own proprietary way of going about it from an architectural standpoint. We also have years and years of production-ready data in different acoustic environments that allows us to fine-tune and just improve the models. By using that intersection of algorithmic as well as data, we find a differentiation. Lastly, I’ll say because we are, you know, our suite is, again, full array. We do them in automotive where you could use the technology without an internet connection, or we use cloud services. We do them in the restaurant space. We do it in enterprises. We’ve had to figure out how to apply our technology into different environments, and it has required us to not just scale at any cost like some of the other big players out there.
We’ve had to figure out how to do this effectively with lower cost. I think that puts constraints, and anytime you want to break through an innovation, constraints really help. I think it’s those types of things that have allowed us to create a really powerful solution that we’re deploying across all our customers.
Tim Fish, Analyst, Piper Sandler Research: Speaking of power and enterprise, you guys made a little news this week. I’m going to have to bug you on that. What does Interactions bring to you? Why Interactions on top of, you know, Amelia? What is complementary here versus an overlap?
Nitesh Sharan, CFO, SoundHound AI: Yeah, so I’m Amelia Feld. Let me touch on strategy first because it’ll fit within that very easily. For those who maybe aren’t as familiar with our story, I’ll say that we, the premise of what we do and the market in which we participate is we believe that now this new era of generative AI, large language models, large reasoning models is breaking through a completely new way of which humans are going to interact with technology. The human-machine interface is completely transforming, and it’s being transformed most notably through natural language conversations. Through all the ways that we interact with technology, we’re used to typing on a QWERTY keyboard. We interact with touch-type swipe on our smartphones. Now natural conversations are taking on more and more activities, and we have different applications I’ll talk through. That is really the ultimate vision.
Within that world, voice AI is the killer app. It’s that tip of the spear that is breaking through. Our view is in, it’s happening in exponential growth right now, but in several years, more and more things will be happening through people just talking to robots and your AI systems at restaurants or your contact center, things where today you’re on hold through an IVR system for a long time, and it’s annoying. That’s getting transformed with better and better natural conversation technology that we’re deploying. Within that, yes, we’ve grown organically substantially over the past several years, and we’ve done some acquisitions, and the most recent, as you noted, was Interactions. They’ve been a pioneer in this space for many years. The company goes back two decades.
They really were early pioneers in intent recognition and had a real augmented solution with technology so that oftentimes when you’re, if you go back and you’re even today, if you call up a cable provider or a large enterprise to change your flights or something, sometimes you’re stuck in that phone tree that everybody hates. Often that’s a way to signal, like, what do you want to do next? Do you have a billing question? Do you have to change your flight? They were pioneers in figuring out intent. Sometimes when you go in and you’re talking, the algorithms have high confidence. You can let the AI continue. When the AI has low confidence, a human could come in and say, wait, they’re trying to return clothes, and it can allow it to go. That was sort of where they got some traction.
They also, over time, because they’re really penetrated well in deep enterprises and iconic brands out there, have come to learn and build their systems around workflow optimization and orchestration. In this next horizon of where the technology is going, that workflow, that optimization of how things get done in enterprise, integrating with enterprise systems, is the real unlock of how we’re going to see this scale faster. Interactions provides great customer base, a great patent portfolio, and again, unique adjacencies to our technology to allow us to scale faster.
Tim Fish, Analyst, Piper Sandler Research: Got it. Especially as a CFO here, I’m going to ask a financial question on it. How big should we think about it? What are you and the team looking to do to extract more value out of it, whether it’s, you know, try to re-accelerate its growth or drive, you know, better profitability, better profile?
Nitesh Sharan, CFO, SoundHound AI: Yeah, I’ll stick to what we announced this week and more to come over time as we announce our earnings. We announced on Tuesday the transaction. The total consideration, we said $60 million upfront, with some earnout potential. A lot of times in the deals we construct, we want to make sure there’s shared interest and their shareholders and our shareholders are winning together. Without giving too much, they’ve been around for a while. They have some great enterprise partners. To your question on what we think we could do, absolutely starts with what’s our growth potential here, both combined in terms of how can we re-accelerate their growth with proper investment. How do we incubate? They have really not only longstanding customers, but great customer relationships. When we went in, we do our diligence and we talk to their customers.
Their customers first say great care and we have a great relationship with them. We’d love to see a little bit more if they could provide omnichannel solutions, for example, which we can bring to bear. There’s opportunity for that cross-sell upsell that we’re excited about. We know we have more to offer these existing customers. I think like any acquisition, you kind of look at cost opportunities, and that’s something we’re certainly working through. These are two large opportunities between that acquisition and prior acquisitions we’ve done that we know there’s some overlap. We’re going to look through that and more to come. We’re still early days of this, just announced, and I’m excited about the opportunity.
Tim Fish, Analyst, Piper Sandler Research: Yeah, it makes sense. Maybe just staying on the conversational AI side of things for a moment. On the Amelia side, they had a big relationship with Nice. Nice just bought Cognigy, so it kind of validates what you guys are doing as well. How do we think about the impacts or what’s coming through the channel or partnerships for Amelia? How is that new avant partnership and contribution longer term going to look like?
Nitesh Sharan, CFO, SoundHound AI: Yeah, I’ll talk about the Nice, Cognigy deal, but before that, even just Amelia, what we were so excited about and what we’ve already seen, and I mentioned this in the last earnings call, that, you know, just 12 months together, nine months together, we’ve been able to reinvest in their go-to-market function and their customer success and the product stack. We’ve already seen the returns in that. We announced some pretty sizable deals, both in terms of cross-sell upsell with the broader portfolio, but also, now getting to your question, sort of opportunity in indirect channel partnerships, of which Avant, Nice, and many others are part of this ecosystem, especially when you get into new environments where there’s specific, you know, regulatory dynamics where an indirect channel partner can be really helpful in a government, for example.
These are opportunities that we continue to look at to try to scale. With respect to their relationship and that opportunity, I think one is validating of the space in the market and the attractiveness and why many people are looking at what combinations make sense for them to grow in this, again, burgeoning new space of conversational AI and voice AI. We think of that as an opportunity and, you know, things we need to be mindful of. We do continue to have a strong relationship with them. We continue to look at expansion opportunities. We think we can be a great partner and provide unique capabilities that even with their acquisition, there’s more that we can offer.
I’d also argue that there’s going to be some other players out there that look at this and go, hey, you know, is there an alternative to Cognigy that maybe they want to look at? They may open some doors. There’s always flavors and changing landscapes. We’re pretty excited with what we have and what we’re, who we’re partnering with.
Tim Fish, Analyst, Piper Sandler Research: Yeah, so on Amelia here, as we think about the growth opportunity and the business model, it does seem as though you guys are seeing more traction with, I’ll say the term subscription side of things, especially with the larger customers. How should we think about that term subscription mix versus kind of the SaaS mix over time here? I think you guys talked about, you know, a pretty decent expansion rate here this past quarter. Is that expansion coming more from upsell of usage or kind of a cross-sell within Amelia?
Nitesh Sharan, CFO, SoundHound AI: Bit of both. It’s really both term and SaaS. Again, just to unpack Amelia, they had three main components to their business. They had the, I’ll call it software, which is both, you know, licensing and SaaS. There’s professional services, and then they had escalation support. What I’ve talked about publicly is on the escalation support. These are really deep penetrated enterprise customers that we’re looking at individual contract by contract to make sure the economic constructs are appropriate, where we need to refine the cost model or, you know, in some case, the contract doesn’t make sense and walk away from, and sometimes we want to do more and grow. That’s continuing. Professional services, that’s an opportunity.
Not only, you know, what do we do day one, but also after we’re in an environment and we’re trying to scale and do more interactions and different types of workflows, that’s an opportunity to continue to grow. On the software side, the mix between whether it’s term or whether it’s SaaS is really often driven by the characteristics of the customer. For example, with certain indirect channel partners that I mentioned we’re growing with, they want to host. They’re already hosting in a cloud environment, so they’re not looking to host further. They want a licensing contract. In that case, if that’s what the customer’s looking for, that’s the direction we’ll go. In certain environments where we talked about a large utility deal a couple of quarters ago, you know, high capital intensity, they pass, you know, kind of more seamless passing of pricing to customers.
Licensing is sort of the, that’s the framework they want. That’s what we’ll work with. The general shift has been overall growth and recurring. Our overall mix of our business has been substantial growth in our recurring model. The last thing I’ll point out, which I think is a shift from this era of cloud that we’ve been living in for 15 years to where this AI, you know, vector is taking us, is more and more of these usage and consumption-based models where a lot of our contracts now embedded sort of interactions. For example, in a hospitality space, if, you know, you get more reservations, that’s incremental economics for us. You see, as long as the technology keeps expanding and the use case grows, then our unit economics continue to grow. That’s kind of the models are typically kind of shifting that way.
Tim Fish, Analyst, Piper Sandler Research: Got it. Maybe shifting gears to literally gears, on the auto side of things. As we think about what that auto vertical, you know, it seems like we’ve been stuck kind of 85, 90 million units a year. You guys actually just landed a sizable customer overseas. How long does it take that size of a deal to sort of roll out in terms of like you getting that initial first order that we could see a second order if their unit volume continues to increase and it goes well? What are you seeing in terms of just the overall production landscape in terms of the adoption of your type of solution versus even the competitive environment like against the Cerence, for example?
Nitesh Sharan, CFO, SoundHound AI: Yeah, I think it depends on the customer. EVs, you can actually deploy really fast. We went, one of the EV partners we’ve talked about is Lucid, and they can, over the waves, deploy across all their full fleet. We’ve actually won more than our fair share with EVs across the globe. I think that is a testament to a couple of things. One is they tend to operate more like a Silicon Valley company. They’re software-centric in the cockpit. Their speed of deployments tends to be much greater. On the traditional guys, there is a journey, and sometimes the model development deployment across their infotainment units could be a multi-year cycle. We’ve seen that shrink, but that’s still the case.
Some of these consumption of these agreements, I would say that, again, sometimes when we’re doing these deals and constructing them, oftentimes we’re trying to build partnerships like a new OEM in China. We’re not trying to come and go. We want to be with them. There’s a multi-year contract. Generally within that, there’s a number of levers in the contract that get negotiated, including is there a minimum commitment, which for us is really important because if we’re going to invest the effort upfront, customize configuration of the technology, certain languages, certain wake words, whatnot, we want to make sure that there’s a commitment on both sides. We ask for that. That does, if it’s an edge part of the deal, all these come with edge and cloud. The cloud is recurring over time.
The edge, you would recognize some amount of it upfront, and the consumption is dependent on unit volume. If unit volumes accelerate, great, that consumption is going to be very short. If for some reason you see tariffs affect the global auto market, like we’ve been seeing this year, that can slow it down a little bit. I will add, though, in the auto space, there’s more. Also, I’ve talked about publicly this ASP expansion, price per unit expansion that we’ve seen. That’s because we continue to expand product sets within it. We got our foray originally with cloud capabilities. We then went into existing customers and added on edge capabilities. We then brought in one of the first companies to bring GenAI into vehicles with our partnership with Stellantis in Europe. That’s incremental price point.
We now have the voice commerce opportunity that we’re adding, and it’s attracting a lot of OEMs along with restaurant partners to kind of say, what else can I do? It is more unit economics. Even though, to your point, to 80 to 85 million global light vehicles—by the way, there’s a heavy truck also opportunity—that can be a stable single-digit grower for a while. We can capture more of it, and we can incrementally add more to those specific units.
Tim Fish, Analyst, Piper Sandler Research: Got it. You mentioned the restaurant, which is actually kind of like one of my favorite parts of your business here. You guys are roughly 14,000 deployed. Is there a way to think about where we are today between Smart Answering versus the drive-thru opportunity? When do we actually start to see that drive-thru uplift in terms of impacting sort of the monthly recurring revenue that we see per restaurant?
Nitesh Sharan, CFO, SoundHound AI: Yeah, so phone ordering is the bulk of our footprint in drive-thru or in restaurants today, partially because it’s much easier to deploy. You know, we have 3,000 location footprint partners who can deploy almost across the full suite. Sometimes there’s an incremental, they have a franchisee they have to, you have to sign an additional contract with. Sometimes we have the corporate MSA that can just deploy down. The phone ordering just scaled much faster. One of the ways we got traction originally was through our channel partner ecosystem. We partnered with Square, Toast, OLO, Oracle, MicroSymphony, et cetera. That allowed us to just penetrate much more quickly. Whereas on the drive-thru side, you have incremental hardware integration requirements. You need to make sure you have the order confirmation board, digital display. You need to work with the right headset partners, microphones.
We do have great partners there like Samsung, Par, HME. One of the ones we’re excited about in Europe is actually led by channel partners. Acrelec brought us the Burger King UK opportunity. I think there’s a lot more opportunity, particularly with the channel ecosystem. We do have customers who come to us, want one throat to choke, and they’re like, can you purchase the hardware to help me deploy it? If it’s right for the customer environment, if the ROI is there, which oftentimes it is, we’re willing to do that. Ultimately, we think it’s this ecosystem that’s building. I think long term, both of the opportunities, both phone and drive-thru, are massive. The TAMs are substantial. It’s a global opportunity. We’re in three continents already. Even just in the domestic market, it’s tremendous. We’re also adding non-ordering capabilities. We have a technology, a product we call Voice Insights.
Now a restaurant doesn’t, if they’re a little hesitant to provide the front-end customer-facing AI, but they really want to understand how their drive-thru system is working, that’s incremental opportunity and product that we sell into them.
Tim Fish, Analyst, Piper Sandler Research: Got it. I think it’s been two or three quarters in a row now that you guys have gotten record number of deployments. How sustainable is this? It’s roughly been about 1,000 every quarter. How sustainable is that? Could we actually see an acceleration behind that?
Nitesh Sharan, CFO, SoundHound AI: I think, yeah, we are, I mean, if you think of, we’re in a lot of businesses with multiple S-curves, I’ll just say. In the restaurant S-curve, we are definitely not in that vertical part yet. To the point on these numbers, when we are talking next year, hopefully at this conference, I hope this will be much, you know, greater numbers because we are, our penetration of the opportunity, again, just in the U.S. alone, 14,000 off of a million food establishments or 300,000 quick service drive-throughs, is a very, very small penetration. You go, what’s our competitive play? Where do we fit in the landscape? We think of the ecosystem as either legacy providers who don’t have the technology to compete or the big tech who have 50 other priorities going on. Even there, they’re not fully invested. We believe we have a really unique moat.
We are accelerating. My hope, although things are never linear in the world, in the world of disruption, things happen episodically, but these numbers should be small compared to where we’re going, and we have a lot more opportunity to capture.
Tim Fish, Analyst, Piper Sandler Research: When you’re talking with customers in this space, my excellent restaurant peer, Brian Mullin, covers the restaurants, and him and I have chatted about this. At the end of the day, same-store sales matters a lot, but it seems like AI is a little bit about cost savings, to a degree. What is the push by customers more on, the cost-saving side, or is it more about enabling that upsellers? Should we think about that?
Nitesh Sharan, CFO, SoundHound AI: Yeah, all of the above, and it changes over time as their own unique situation kind of evolves. Each customer, sometimes it starts with like a cost savings because commodity costs are going up or it’s, you know, a little labor arbitrage because they can’t hire the staff. Now the AI can supplement. By the way, some of our restaurants are 24/7, very hard to hire for people to do that night shift. It is definitely cost. It’s labor, but it’s also speed of service. The algorithms can be dialed up. If you are in a slower period and you want to say, would you like to supersize that or would you like dessert with that? You can do that. Maybe in a busy lunch hour, speed is super critical. You could dial that back. It’s really around cost and, you know, revenue uplift and just accuracy.
Like how effective is this? Is this a place where, you know, customers want to go and, you know, they’re excited about the customer interaction. It allows to free up the humans in the store to actually work on customer service, hand you your food with a smile versus really chasing around everything.
Tim Fish, Analyst, Piper Sandler Research: Yeah, so you mentioned actually voice commerce before, and it’s kind of an interesting bridge between two of your big opportunities here, right? Between auto and restaurant. Think about what you’re seeing in terms of that early traction, especially with the OEMs, and sort of what I’ll say commission structure it would look like for you guys.
Nitesh Sharan, CFO, SoundHound AI: It’s really attracted a lot of new OEMs that weren’t maybe our direct voice assistant partners as well, our current partners we’re talking to. It’s bringing in a lot of restaurants that also may not have been our current subscription restaurant customers. It’s this new opportunity where restaurants are like, great, I’m going to get new leads, and automotive companies are like, great, I can make more revenue. It’s a huge opportunity, win-win, and ultimately the big benefit is for the consumer who can actually pick up their coffee on their way to work seamlessly. The economic model right now is sort of, we’re doing proof of concept, but the idea in the spirit of it is really an ecosystem where everybody wins.
The idea of a convenience fee that can be shared across, if you consummate a cappuccino on the way to work, and there’s a convenience fee associated, we would share some of that economics with the car manufacturer. The way we try to distinguish it is if an existing voice assistant car manufacturer that we’re already working with, there’s a greater share of economics. If not, we still want you to participate. We’ll attract more people. There’s a bit of that. Over time, there’s an advertising element that could play into this. The economic model is kind of coming together, but the idea is that the product manufacturers, including car manufacturers or TV manufacturers or other product devices, are going to see a whole new thread of revenue generation opportunity for them.
The restaurant locations who are always clamoring for new demand and new consumers are going to get new leads. It’s a huge opportunity for them and ultimately the consumer benefits.
Tim Fish, Analyst, Piper Sandler Research: Got it. With remaining a couple of minutes here, especially as a person in your chair, a lot of questions we get are around the second half numbers. You guys outperformed in Q2, changed the guide a little bit upwards here, but you guys are one of the few companies that are organically what we would call hypergrowthy still. How do we think about what you have to do for, I’ll call it go-gets for the second half of the year to hit numbers, what kind of visibility you have, and what do you think we actually grow? I know you’re not going to guide here for next year at this point, but what sort of CAGR should we think about for the next few years for SoundHound AI?
Nitesh Sharan, CFO, SoundHound AI: We’re an early stage disruptor in a massive market, and it’s a market that’s just burgeoning. If I go to where I started around conversational AI being the next major inflection of how humans interact with technology and voice AI is that spearhead or that killer app in that world, our penetration level, whether in auto or in restaurants or in the enterprise sector, is really small, and we just need to grow with our existing customers. I’ve previously talked about our bookings backlog numbers being north of $1 billion and huge opportunity just to penetrate deeper within our customer base. The other side of it is a lot of these enterprise customers we’re working with are multiples of the size. I mean, they’re much, much bigger than us. For them, like a small contract is a big deal for our quarter-to-quarter numbers.
That’s why it becomes lumpy for us because we have a massive pipeline, and we’re always working through which deal makes the most sense with our limited resources. Where can we be successful? Those are the trade-offs we’re going through every day. We think there’s a lot of upside just with the existing customer base. We think there’s a lot more that we’re going after that are in a pipeline, late-stage qualified pipeline. Frankly, opportunities like this acquisition just add to it because now it extends our opportunity. It allows us to bring new products, upsell opportunities. It penetrates us into new customer bases. That’s what gets us excited about going ahead. Long term, I’ve said this before. I mean, we are no reason we shouldn’t be a strong double-digit grower in perpetuity, maybe not perpetuity, but for a very long time because we’re just getting going.
Tim Fish, Analyst, Piper Sandler Research: Yeah, you guys have a ton of opportunities, and you guys are executing on it so far. I really appreciate you coming. I think that’s a good spot to end. Thank you everybody for joining us.
Nitesh Sharan, CFO, SoundHound AI: Thanks for having me.
This article was generated with the support of AI and reviewed by an editor. For more information see our T&C.