SOUNDHOUND AI, INC. 10-K

Item 1. Business Our Mission SoundHound’s mission is to voice-enable the world with conversational intelligence through an independent AI platform enabling humans to interact with products and services like they interact with each other — by speaking naturally. Company Overview SoundHound is a leading innovator of conversational intelligence, offering an independent Voice AI platform that enables businesses across industries to deliver high quality conversational experiences to their customers. Built on proprietary Speech-to-Meaning ® and Deep Meaning Understanding ® technologies developed over the past 17 years, our advanced Voice AI platform provides exceptional speed and accuracy. We envision a future where people interact with products on a daily basis through voice-enabled AI. Our technology provides a conversational voice-enabled user interface, allowing a more natural and more productive way of interacting with the products and services we use. SoundHound is also committed to enabling product creators to design, customize, differentiate, innovate, and monetize the voice interfaces to their own products and services, as opposed to outsourcing to a third-party assistant. More often than not, we currently interact with voice-enabled third-party assistants using halted speech patterns, consciously dividing queries into limited, broken, and unnatural instructions. By contrast, using SoundHound, businesses can voice-enable their products so consumers can say things like, “Turn off the air conditioning and lower the windows,” while in their cars, “Find romantic comedies released in the last year,” while streaming on their TV, and even place food orders before arriving at a restaurant by talking to their cars, TVs, or other “internet of things” (“IoT”) devices. Additionally, SoundHound’s technology can address complex user queries such as, “Show me all restaurants within half a mile of the Space Needle that are open past 9pm on Wednesdays and have outdoor seating,” and follow-on qualifications such as “Okay, don’t show me anything with less than 3 stars or fast food.” SoundHound’s technology is currently being used globally by customers that include Hyundai, Mercedes-Benz, Pandora, Snap, VIZIO, Square, Toast, Oracle, KIA, and Stellantis. We have seen significant inflection in customer adoption of our technology, as measured through activity on our Houndify platform, which surpassed 1 billion annual queries in 2021 and experienced over 85% growth during 2022. Our current customers, which in many cases have contractual obligations that average multiple years (often between three to five years), span multiple industries and geographies, and together have an estimated combined reach of over two billion end users. We consider our long-term customers our “partners” and, by growing these strategic relationships, SoundHound’s revenues have grown by more than 45% year over year each fiscal year over the past four years. We support our customers by providing them access to Houndify ® — an open-access platform that allows developers to leverage SoundHound’s Voice AI technology and a library of over 100 content domains, including commonly used domains for points of interest, weather, flight status, sports, and more. To ensure that our content domains continue to evolve and grow, our platform is built on our breakthrough Collective AI ® — an architecture for connecting domain knowledge, which encourages collaboration and contribution among developers, is always learning, and is greater than the sum of its parts. This architecture allows us to improve the understanding capability of our platform super-linearly and even exponentially based on linear contributions because of how the domains interact with one another. They can be inter-connected and can learn from each other and, as developers contribute to the platform, its understanding capability can grow exponentially. Our technology is backed by SoundHound’s investments in intellectual property, with over 120 patents granted and over 140 patents pending, spanning multiple fields including speech recognition, natural language, machine learning, monetization, and more. We have achieved this critical momentum in part thanks to a leadership team with deep expertise and proven ability to attract and retain talent. 1 Table of Contents Market Opportunity A significant addressable market To take advantage of opportunities created by the expansion of Voice AI, companies across industries are turning to independent providers with disruptive technology offerings to help them design and build custom voice assistants. The demand for Voice AI technology is growing exponentially as companies compete to lead the market with the best technology, excellent customer relationships, and greatest innovations. Industry sources predict that 90% of new vehicles globally will have voice assistants by 2028 and that there will be 75 billion connected devices worldwide by 2025. The number of devices with their own voice assistant is expected to surpass the world’s population within four years. Across industries, 94% of large companies expect to use Voice AI within two years, according to a Pindrop Security, Inc. study, indicating the imperative most companies feel to provide a voice user interface to their products and services. Custom, branded voice assistants are allowing companies to improve their customer interactions at every touchpoint. In a recent study by Opus Research (sponsored by SoundHound), 77% of business leaders across eight industries stated that voice assistants drive value through improved customer support and 76% agreed that they help companies to control their brand identity and user experience. For those already employing a voice assistant, 81% are using the technology to improve customer service and 52% are using them for sales activity. For the estimated 75 billion IoT devices in existence by 2025, we believe many of them won’t have the physical or economical room to add a keyboard or touch screen, but they can easily have a small, inexpensive microphone. With a microphone, companies can add Voice AI to their products, unleashing the full capabilities of the internet to their end users. Voice AI can bring these products “to life”, and convert IoT to AI IoT. According to Juniper Research, the broader market for Voice AI transactions will exceed $160 billion by 2026, enabled by a growing need for Voice AI in many industries. All these trends are adding up to exponential growth for Voice AI in a variety of markets, including IoT, automotive, retail, hospitality, enterprise, healthcare, contact center, and banking and finance. Strategy We believe that Voice AI is poised to be the next disruption in computing and the timing is right for our company and our vision. Human-computer interfaces became popularized through computers offering a keyboard and mouse. After that came mobile devices with touch screens. We now have IoT devices with Voice AI. After the internet became mainstream, every company needed to have a website. After the mobile ecosystem became mature, every company needed to have a mobile strategy. We note that some very successful companies were created purely on the mobile ecosystem. We believe the same concept will apply to the world of Voice AI where every company will need to have a strategy and there will exist success stories built on top of platforms like ours. These interfaces will co-exist: for example, mobile and touch screens did not completely replace computers with a keyboard and mouse, therefore Voice AI will also co-exist with computers and mobile. However, with billions of IoT devices that don’t have a keyboard or a mouse or a touch screen, voice will be the preferred or the only way to interact with these devices. Our market position is strengthened by the significant technical barriers to entry that often face potential participants in the Voice AI space, reducing the likelihood that significant competitors will enter the market in the near term. Technical Development In 2005, when, as recent graduates from Stanford, our founders began envisioning a future where we would be talking to products just as we do other people. Inspired to make their vision a reality, the Company was founded with a simple and powerful mission: Add Voice AI to everything. To this day, the Company is still run by its original founding team. 2 Table of Contents SoundHound’s technology represents the evolution of several disruptive breakthroughs in voice-AI and sound recognition developed over the past 15+ years. Using innovative audio and music identification technology, in 2009, our founders launched the SoundHound music identification app. Since its inception, the app has received over 315 million downloads. In parallel, the founding team worked in stealth on a fundamentally new technology — one that had the potential to revolutionize human-to-computer voice interaction. Up until then, all voice interactions relied on speech-to-text and then text-to-meaning. Our engineers knew that for a true voice engine to flourish, it needed to understand speech directly, just like humans do. In 2015, SoundHound introduced the Houndify Voice AI platform incorporating breakthrough Speech-to-Meaning and Deep Meaning Understanding technologies, which, to SoundHound’s knowledge represented voice interaction technologies not yet broadly available at the time as its foundation. In 2022, SoundHound introduced Dynamic Interaction™, a category-level breakthrough in conversational AI that raises the bar for human-computer interaction by not only recognizing and understanding speech, but also responding and acting in real-time. Where existing voice technology requires wake words and relies on turn-taking with awkward pauses to process requests, Dynamic Interaction uses the twin technologies of fragment parsing – which breaks speech down to partial-utterances and processes them in real-time – and full-duplex audio-visual integration to create an instantaneous, next-generation experience in human-computer interaction. Additionally, in March 2023, we unveiled a new product called "SoundHound Chat AI", which we believe will usher in a new phase of voice-enabled, conversational AI by combining the power of software engineering and machine learning generative AI. SoundHound’s Vision A typical reaction to SoundHound’s technology is “Wow! Even humans cannot do this.” This is the crux of our vision: Create a Voice AI platform that exceeds human capabilities and brings value and delight via an ecosystem of billions of products enhanced by innovation and monetization opportunities. Computers are better than humans at computing, but they’re not always better than humans at performing certain tasks. For example, we used to easily beat computers in the game of chess, but now computers can even top grandmasters. When it comes to language understanding, the general perception is computers are not yet as good as humans. People talk using complex conversations with each other, but when they talk to voice assistants, they lower their expectations and use short, simple, keyword-based queries. We aim to change that by making computers better than humans in language understanding. As a result, SoundHound can make humans more productive and help make the world a better place. To provide another example, if a user is looking for a restaurant, but anything other than Chinese (because they had Chinese food yesterday) they can ask the concierge of a hotel: “Show me restaurants, excluding Chinese”, and the concierge can point the customer in the right direction. But when talking to a typical voice assistant, users often don’t ask the question: “show me restaurants, excluding Chinese”, because they don’t expect that it will work. And in fact, most other voice assistants will offer Chinese restaurants when asked to show restaurants excluding Chinese. With SoundHound’s technology, users can ask even more complicated questions, such as: “show me Asian restaurants in San Francisco, excluding Chinese and Japanese, and only show the ones that have more than three stars and are open after 9:00 PM on Wednesdays.” In addition, users can follow up and ask it to refine the criteria such as “remove Korean and Vietnamese, sort by rating then by price, and only show the ones that are good for kids and have a patio.” Additionally, through SoundHound's AI-driven restaurant voice ordering solution, SoundHound for Restaurants, automated voice technology can be deployed by any restaurant to take food and drink orders over the phone, via menu kiosks or at the drive-thru. Its sophisticated technology learns the menu of each business and can answer questions, accept modifications and even upsell, helping the restaurant process more orders quickly and efficiently. Our vision further includes empowering billions of devices around the world using our technology, with innovation and monetization opportunities for the product creators that integrate the Houndify platform. It means product creators can not only use Voice AI to make their product better they can also generate incremental recurring revenues from customer interactions. 3 Table of Contents Our vision also places a high emphasis on user experience. Before monetization growth can occur, delivering value and delight to end users is paramount. As a result, the most effective monetizable interactions will be those that flow naturally based on context, create value for the end-user, and would not be perceived by users as intrusive advertisements. We note that today’s voice assistants are far from being able to answer every question and perform every task, and our vision is long-term. However, we have already achieved superhuman capabilities in many of the domains that we have delivered to our customers. Technology Breakthroughs SoundHound’s momentum in the Voice AI market can, in large part, be attributed to our large number of technology breakthroughs. These are anchored by three important innovations: Speech-to-Meaning, Deep Meaning Understanding and Collective AI. Speech-to-Meaning Speech-to-Meaning refers to SoundHound’s ability to convert speech to meaning simultaneously and in real time. Most traditional approaches first convert speech to text, and then convert text to meaning. This approach can be both slower and less accurate. It’s slower because the two steps are done in sequence, and the additional processing time of the second step can be noticeable by the end user. It can also be less accurate because if the first step of speech to text makes a mistake, the resulting incorrect text is then sent to the second step, and the error further propagates. Our development of Speech to Meaning technology was inspired by the human brain. As we listen to someone speaking, our brain does not convert speech to text, and then text to meaning. Instead, our brain converts speech to meaning simultaneously and in real time. With Speech-to-Meaning, as you speak to SoundHound’s technology, it performs both speech recognition and language understanding simultaneously, which results in faster response time and higher accuracy, because real-time language understanding can feed into the real-time speech recognizer as additional information to reduce errors. Deep Meaning Understanding Deep Meaning Understanding is our innovative approach to language understanding that allows our Voice AI platform to understand highly complex conversation. 4 Table of Contents For example it can understand: “Show me hotels in San Francisco that are less than $600, but not less than $300, are pet friendly, have a gym and a pool with at least three stars staying for two nights, and don’t include anything that doesn’t have Wi-Fi.” A complex search like this will take many minutes to perform on a website with complex forms, but it can be done within a few seconds using SoundHound technology, which we believe to be unique in its ability to handle complex queries of this nature at scale. Collective AI The third breakthrough is called Collective AI: an architecture that gives potential to SoundHound to improve the understanding capability of its platform exponentially based on linear contributions Most other platforms add skills or domains that are separate and don’t interact with each other. For them, linear contribution results in linear growth in understanding, which is less scalable. With the Collective AI architecture, SoundHound domains can be interconnected and learn from each other. As developers contribute to the platform, the platform’s understanding capability can grow exponentially. SoundHound Products Houndify platform SoundHound’s Voice AI platform combines advanced AI with engineering expertise to help brands build conversational voice assistants. From proprietary components to customizable and scalable solutions, we offer tools to build a highly accurate and responsive voice user interface. The suite of Houndify tools includes Application Programming Interfaces (“API”) for text and voice queries, support for custom commands, extensive library of content domains, inclusive Software Development Kit platforms, collaboration capabilities, diagnostic tools, and built-in analytics. Houndify provides a web API that takes in text queries or audio and returns actionable JavaScript Object Notation to anyone with an internet connection wanting to add Voice AI to any product or application. 5 Table of Contents Automatic Speech Recognition (“ASR”) Our highly optimized, tunable, and scalable ASR engine supports vocabulary sizes containing millions of words. Houndify’s machine learning infrastructure allows us to tune the engine to achieve optimal Computer Processing Unit (“CPU”) performance while delivering high accuracy rates. Houndify’s language and acoustic modelling architecture also uses machine learning to increase word recognition accuracy. Rapid iteration is possible due to our accelerated training pipeline and architecture that improves as data is collected. Highly accurate transcriptions result from advanced acoustic models trained to perform in a variety of scenarios — including in severely noisy environments and when accented language is spoken. Natural Language Understanding (“NLU”) Our proprietary Speech-to-Meaning technology tracks speech in real-time and understands the context, even before the user has finished speaking. Instead of the typical two-step process of transcribing speech into text and then passing the text into an NLU model, Houndify can accomplish both of these tasks in one step, delivering faster and more accurate results. Houndify’s ability to process and understand speech the instant a user stops speaking gives voice assistants the ability to respond faster. Understanding speech in real-time without requiring additional processing or waiting for the user to finish speaking creates responsive and natural conversations between people and products. By understanding context, Houndify responds accurately to users by distinguishing between similar words and names. Our NLU can discern the difference between words that sound the same, but have different spellings and meanings. For example, if users want to navigate to 272 Hoch Street in Dayton, Ohio, it won’t look for Hawk Street. Using our proprietary Deep Meaning Understanding technology, a custom voice assistant can handle complex queries with compound criteria including conversational follow up, address multiple questions and filter results simultaneously — accurately and quickly answering users’ most complex questions. Wake Words Wake words are the entry point into branded voice experiences, allowing users to invoke the assistant by literally speaking the company’s name. Examples range from “Hey Pandora” in a mobile app to “Hey Peugeot” within a vehicle. Rigorous development and testing enable our wake words to perform in noisy environments and minimize false-positives or false-negatives. We use advanced machine learning algorithms and Deep Neural Networks to provide broad robustness to our high-volume training data, resulting in high accuracy. Custom Domains Our library of over 100 public domains is available with a free Houndify account. Houndify public domains give developers instant access to a broad range of content to fit their unique use cases. This includes multi-category content intended to appeal to broad range of audiences, including, for instance, sports scores, weather, podcasts, travel information, recipes, stock prices, among many others. Companies can enhance product functionality or proprietary operations with Houndify Private Domains, allowing customization and development of more specific content. Customers who subscribe for this service have full access to their private domains securely on the Houndify platform while retaining the ability to iterate and update content. For example, an automotive manufacturer can make helpful updates about the car’s user manual over time. In this way, SoundHound becomes a long-term “partner” to its customers, helping companies create the domains that they need in order to improve brand value for their own customers or end users. Text-to-speech (“TTS”) A high quality TTS helps companies create a unique voice that differentiates them from the competition. Brands can fully express their personality by choosing the gender, tone, and personality that will become their vocal identity. 6 Table of Contents Our machine learning algorithms transform recorded voices into large databases of spoken sounds to form entire vocabularies of natural language — adapted to the user’s environment. We can transform any voice to generate a high-quality TTS with a small CPU footprint. Edge (Embedded) A fully-embedded voice solution for brands seeking the convenience of a voice user interface without the privacy or connectivity concerns of the internet. Includes full access to custom commands and the ability to instantly update commands during development. To harness the capabilities of full cloud connectivity with the reliability of embedded voice technology. Houndify Edge Hybrid solutions are designed to ensure that devices are always-on and responsive to commands. Allows for over-the-air product updates and a broader voice experience with the level of cloud-connectivity that best matches the product and its users. Competitive Strengths and Differentiators Independent Disruptor We believe there is a growing void and demand for an independent Voice AI platform such as SoundHound. Looking across industries, there has been a notable rise of independent disruptors. These new companies were able to beat the legacy giants using focused technologies and business models. The majority of participants in the Voice AI industry can be characterized as either “big tech” companies (meaning large corporations offering Voice AI as an extension of other services) and “legacy vendors” (meaning long-time Voice-AI industry participants with older technologies). Voice AI offerings from big tech companies are primarily an extension of their services. In many cases, rather than strengthening or interacting with the host product, big tech offerings tend to take over the entire product experience, thus “hijacking” the company’s brand, users, and data. As a result, brands relying on big tech Voice AI frequently experience decreased ability to innovate, differentiate, and customize the way that their products interact with Voice AI platforms and/or end users. In some cases, these big tech Voice AI providers even compete with the customers whose products their technologies support, making them increasingly less attractive as a choice for a voice interface. The Voice AI market currently also includes “legacy vendors,” a number of whom offer dated technologies at a high price. Furthermore, in many cases, legacy vendor technologies still require significant effort by product creators to turn legacy AI product offerings into solutions that can compete with the quality of Voice AI products currently offered by big tech companies, which is oftentimes neither economical nor practical. Due to the high barrier to entry in Voice AI, the number of full solution platform providers is very limited. In our view, it takes many years and significant investment to build all the components of Voice AI. It then takes further time and resources to make the solutions competitive, mature, and viable for adoption. Finally, it requires significant investment to globalize the solution in multiple languages and regions. Although the number of platforms is limited, we note that the big tech players have significant resources. Against this landscape, SoundHound has achieved its successes to date through technology innovation, business model innovation, and global alliances that we believe make us well-positioned to continue to grow and execute our business plan. 7 Table of Contents We view the current environment as an opportunity to provide disruptive technologies with capabilities we believe are superior to existing alternatives, allowing customers to maintain their brand, control the user experience, get access to the data, and define their own privacy policies, while being able to customize, differentiate, innovate, and monetize. Revenue Model Market Momentum Our entry into the Voice AI space began with 10 years of constant innovation in “stealth” building disruptive technologies in Voice AI using innovative approaches. Our goal was to build a differentiated Voice AI technology that we fully own and which is significantly better than other solutions in the market. We achieved that goal and unveiled the result in 2015, launching it as the Houndify platform in 2016. Building a Diverse, Global Customer Base SoundHound continues to expand the capabilities that make us well-positioned to serve the needs of customers globally. We have grown our solution from a single language capability to 25 languages. Our customers include a range of product and service providers of all sizes, spanning a range of industries, including automotive, IoT, apps, restaurants, and more. Many of our global customers have end users in multiple regions and industries and we have seen our products successfully used by them across multiple contexts and purposes. Three pillars for growth We have identified three pillars for revenue growth: Royalties, Subscription, and Monetization, and all three pillars contribute to our current revenues today. While the majority of current revenues come from royalties, over time we expect 8 Table of Contents our revenues from subscription and monetization pillars to increase and eventually represent a larger portion of our overall revenues. (1) Royalties : This involves voice-enabling a product. The product creator pays us a royalty based on volume, usage, or duration. SoundHound collects royalties when Houndify is integrated into a product such as a car, smart speaker, or appliance. (2) Subscription : This involves voice-enabling a service that doesn’t rely on a physical product. Examples include when SoundHound enables customer service or food ordering for restaurants or content management, appointments, or voice commerce, we generate subscription revenue from the service providers. (3) Monetization : This pillar creates an ecosystem that enables monetization services in products and services from both pillar one and pillar two. When users of a voice-enabled product access the voice-enabled monetization, this creates new leads and transactions. SoundHound generates monetization revenue for generating these leads and transactions, and will share revenue with the product creators. Through December 31, 2022, we have not generated revenue from leads and transactions on voice-enabled products from voice-enabled services other than from the SoundHound AI music identification app. Going forward, SoundHound expects monetization revenue to be generated through a combination of advertising revenue from the music identification app and from leads and transactions on voice-enabled products from voice-enabled services. For example, imagine a driver asks their Houndify-enabled car what restaurants are on the way home. The car can then respond with options, including restaurants that are also on the Houndify platform. These restaurants can accept orders directly from the car and the user can order using their voice. In that case, we will have unlocked a seamless transaction. Accordingly, the restaurant will pay us for that order and we will share that revenue with the car manufacturer. In this example, each party receives value in the ecosystem. The restaurant gets value because they generated a new lead and booked a sale. The driver gets value because they have placed an order just by speaking to their car. And finally, the car manufacturer delivered value to the end user and generated additional revenue from the usage of their product. We expect the disruptive three-pillar business model will create a monetization flywheel. As more products integrate into our platform, more users will use it, and more services will choose to integrate as well. This creates even more usage, 9 Table of Contents and results in a flow of revenue share to product creators, which further encourages even greater adoption and integration with our platform, and the cycle will perpetually continue and expand. This ecosystem has a compound impact on our business. First, it increases adoption because more products will be motivated to integrate into our platform. Without the three-pillar model, only products that can afford the cost of Voice AI would be able to adopt it. With this model, products will be able to see a path to add incremental recurring revenue from the usage of their product, increasing overall return on investment while making their product better, which will increase overall adoption. Second, our total available market increases with new revenue streams and puts SoundHound on a trajectory with much higher potential. SoundHound’s undeniable criteria for adoption When it comes to criteria for adoption, our goal is to win on every dimension. We believe our product will be adopted for the following reasons: 10 Table of Contents The first two criteria customers typically consider are technology and brand control. We strive to provide our customers with the best technology, and we provide a white label solution giving our customers control of their brands. In some industries you may have to choose between technology and brand control. In our case, we offer our customers the best of both, we enable them to offer disruptive technologies to their users while maintaining control of their brand and user experience. With our disruptive monetization strategy, we also provide a path to monetization. By choosing our platform, product creators can generate additional revenue while making their product better using Voice AI, providing further incentive to choose our platform. We offer a superior ecosystem with Collective AI along with definable privacy controls, which are becoming increasingly important in the industry of Voice AI. Additionally, there is no conflict of interest between us and our partners and customers as we do not compete with them (as some other Voice AI vendors do). We also offer edge and hybrid solutions. This means our technology can optionally run without a cloud connection for increased flexibility and privacy. Our focus is on delivering the most advanced Voice AI in the world and thus allowing our partners to differentiate and innovate their overall experiences for their brands. We strongly believe that product creators know their product and users best. The idea of a single third-party assistant taking over their product is not reflective of our anticipated future. We envision that every product will have its own identity, and they will have Voice AI customized in different ways. They can each tap into a single Collective AI to access the ever-growing set of domains, but the product creators can innovate on top of Collective AI and create value for the end users in their own way. Research and Development Research and development have been the foundation of our company since its origin. We have invested significantly over the years in making conversational Voice AI technology by leveraging advanced technologies across signal processing, speech recognition, machine learning and more. The complexities of our design and the associated technological breakthroughs has required more than ten years of research and development activities to fine tune our technology for commercial use. We continue to explore different innovation strategies to strengthen our capabilities. In particular, we are continually working to improve the speed and accuracy of our Voice AI solutions and make customer adoption easier and faster. We have protected our innovations throughout with patents and trade secrets, and we have continued to strengthen our competitive positioning by staying on top of the latest advancements in the Voice AI industry. We expect to continue to keep research and development and, more broadly, innovation and product quality at the forefront of our strategy and core focus. To further our transformative strategy as a leading innovator, we expect research and development to remain a significant part of our growth and investment. Our research and development activities are centered in our Santa Clara headquarters, but we also carry out research and development in other locations, including internationally. Our employees’ expertise ranges from cloud-based software development to embedded hardware integration. Sales and Marketing We take an insight-driven, account-based marketing approach to build and expand our relationships with commercial partners. We collect feedback directly from commercial partners to garner insights that help drive the business and product. We also work with analysts and higher education institutions to conduct studies, test and validate technology performance, providing key proof points for commercial partners considering our products. In parallel, marketing and communications drive our brand equity and narrative through ongoing announcements, campaigns, events, speaking opportunities, and public relations efforts. Our demand generation efforts span the full customer funnel to target prospects across a variety of channels including: advertising, email, social media, and search engines. We employ Account-Based Marketing (“ABM”) leveraging leading-edge platforms supported by fully-current Marketing Automation tools to capture and nurture business leads through to Marketing Qualified Leads (“MQL”), Sales Accepted Leads (“SAL”), and Sales Qualified Leads (“SQL”) ultimately to drive towards return-on-investment-positive marketing expenditures. 11 Table of Contents Sales and marketing will play a critical role in the next phase of our evolution as a company, with key ongoing investments in our team and leadership. While our products are already scaling with existing customers, we see significant opportunities to grow into new markets and verticals. Increased sales and marketing efforts will enable us to capitalize on the tremendous momentum we are building and we expect to continue expanding resources to grow our personnel and leadership team focused on sales and marketing. As part of a targeted restructuring, the Company will increase its focus on SoundHound for Restaurants while reducing investment in new verticals and already completed language development projects. It will maintain its growing licensing business in smart devices, TV and automotive verticals. Key Strategic Global Partners SoundHound sees strategic partnerships as the foundation for our ongoing growth and success. Our deep collaboration with leading companies across industries has allowed our technology to reach millions of customers. Largely through our existing customer base, query volume exceeded 1 billion at the end of 2021 and experienced over 85% growth during 2022. Information provided to us by our strategic partners suggest that our customers’ products have a reach of over a billion users. Some of the strategic partners that are featured here utilizes SoundHound’s product integrations: • Mercedes: MBUX (multiple models in North America) • Hyundai/KIA: Dynamic Voice Recognition (multiple models in North America and India) • Honda: Select models in Japan and Europe • Snapchat: Voice Scan and Captioning features • Pandora: In-app Voice Mode • VIZIO: Smart TV Voice Remote for multiple product lines • White Castle: Voice AI Drive Through • Stellantis N.V.: Multiple models in Europe Intellectual Property SoundHound’s intellectual property portfolio includes over 120 patents granted and over 140 patents pending. These patents cover areas such as speech recognition, natural language understanding, machine learning, human interfaces, and others, including monetization and advertising. Out of our over 260 patents granted and pending, more than 40 of these patents are in conversational monetization. Because we predict that search traffic will change from keyword-based queries to conversational interactions, we have a large number of patents in the area of conversational advertising. Human Capital At SoundHound, our success is driven by the quality of our people, who we believe are among the best and brightest in the industry. We strive to create an environment that supports employee success and a culture where everyone has a voice in driving innovation. Employees As of December 31, 2022, we had approximately 430 full-time employees, of which approximately 57% were in the United States and 43% were in our international locations. Our restructuring efforts in January 2023 included a reduction of our then-current workforce by approximately 40%. We have not experienced work stoppages and believe our employee relations are good. 12 Table of Contents Employee Experience Emerging from the COVID-19 pandemic, we have further increased our focus on delivering a great employee experience, anchored by meaningful work, career opportunities and well-being, to continue to attract and retain high quality talent. We support employee well-being from multiple dimensions, including economic, health, development, and lifestyle with a comprehensive suite of compensation, benefits and flexible work options. For the vast majority of SoundHound’s employees, productivity is not tied to being in an office and collaboration can happen between people anywhere. SoundHound’s modern approach to work gives our employees more flexibility to choose where and how to work. Depending on their role, many of our employees can choose their office location, as well as work from home some or all of the time. We believe this allows us to engage with a wider pool of talent and retain employees who want or need more flexibility. Diversity and Inclusion We recognize that everyone deserves respect and equal treatment, regardless of gender, race, ethnicity, age, disability, sexual orientation, gender identity, cultural background or religious belief. We are creating an inclusive workplace through community-building, training and internal awareness. In 2022, we continued to support our many community partners and employee resource groups that build community for employees from underrepresented groups. Government Regulations We are subject to various laws, regulations, and permitting requirements of federal, state, and local authorities, related to health and safety, consumer privacy, anti-corruption and export controls. The foregoing may include the U.S. Foreign Corrupt Practices Act of 1977, the U.S. Export Administration Regulations, Money Laundering Control Act of 1986 and any other equivalent or comparable laws of other countries. We believe that we are in material compliance with all such laws, regulations, and permitting requirements. Legal Proceedings From time to time, we may become involved in actions, claims, suits and other legal proceedings arising in the ordinary course of our business, including assertions by third parties relating to personal injuries sustained using our products and services, intellectual property infringement, breaches of contract or warranties or employment-related matters. We are not currently a party to any actions, claims, suits or other legal proceedings the outcome of which, if determined adversely to us, would individually or in the aggregate have a material adverse effect on our business, financial condition and results of operations. Available Information Our website is located at www.soundhound.com , and our investor relations website is located at investors.soundhound.com . Access to copies of our SEC filings, corporate governance information, and other items that may be material or of interest to our investors is available via our investor relations website. The contents of our websites are not incorporated by reference into this Annual Report on Form 10-K or in any other report or document we file with the SEC, and any references to our websites are intended to be inactive textual references only.