Long drive-through queues and the hassle of ordering food on the go are problems familiar to many drivers. SoundHound AI, a voice artificial intelligence (voice AI) company, addresses these with its first in-vehicle voice commerce ecosystem that enables drivers to place food orders entirely through voice commands. The platform was unveiled at the Consumer Electronics Show (CES) in Las Vegas in January 2025 and at the National Restaurant Association Show in Chicago in May 2025.
Founded as a music recognition app and later rebranded, SoundHound announced in November 2021 that it would go public through a SPAC merger with Archimedes Tech SPAC Partners Co. – valuing the company at $2.1 billion. The company officially began trading on Nasdaq on April 28, 2022. This move gave SoundHound the capital and visibility needed to accelerate its commercial offerings, one of which is its in-car voice ordering system.
The system operates through a three-step framework: restaurant discovery, voice ordering, and convenient pickup. Unlike standard search engines, restaurant discovery is enhanced by personalized recommendations. The voice ordering step is powered by their Speech-to-Meaning® and Deep Meaning Understanding® technologies, designed to process orders efficiently. Food pickup can be synchronized with the vehicle’s GPS navigation, ensuring food freshness and customer convenience.
How Voice AI Is Powering the Next $98 Billion Opportunity
Voice AI is now central to enterprise strategies and consumer products, expanding from mobile phones to commerce, customer service, and connected devices. Innovations like Samsung’s home robot Ballie and Deepgram’s Voice Agent API reflect this shift. SoundHound AI, originally launched as the music recognition app Midomi, evolved into a leading voice AI company with proprietary Speech-to-Meaning® and Deep Meaning Understanding® technologies. Its Houndify platform powers scalable applications, including in-car voice assistants for navigation, communication, and transactions.
In-car voice platform
Beyond simply responding to commands, SoundHound AI’s in-car voice commerce platform exemplifies how voice AI can directly drive revenue generation and brand engagement. According to the company, this platform presents a $35 billion annual revenue opportunity for automakers, considering factors like projected subscription, transaction, data brokerage, and advertising revenue. Furthermore, voice AI integrations are opportunities for increasing car brand loyalty as these enhanced features can help a brand stand out from others that only deliver traditional in-vehicle controls. Software becomes a critical business driver for cars as they evolve from being a mere mode of transportation into a commerce-enabled platform.
Voice AI in the food industry
For the fast food industry, the platform could unlock $63 billion in added revenue, driven by increased order volume, faster service, and personalized upselling. According to a commissioned study among 1210 U.S. drivers, 77% of the drivers who regularly place food orders while on the go would rather place their orders using an AI voice assistant than go through a traditional drive-thru queue. Aside from being more efficient, in-vehicle voice ordering is also viewed by 63% of the respondents as a safer alternative due to the lessened hand contact involved in comparison to usual on-the-go purchasing methods. These figures point to a major shift in how restaurants must engage with customers, with voice AI serving as a tool to meet evolving consumer preferences.
Before launching its in-car voice commerce platform, SoundHound AI had already ventured into food ordering innovations. In 2020, SoundHound AI partnered with Mastercard to co-develop voice-enabled drive-through systems for fast food restaurants. This partnership combined SoundHound AI’s Houndify platform with Mastercard’s commerce infrastructure to enable touchless transactions, which was a timely innovation during the pandemic. In June 2024, SoundHound AI acquired Allset, a mobile food ordering platform, further extending its reach in food commerce and facilitating integration of its voice technology into both vehicles and restaurants.
SoundHound AI: Patenting Activity
A close look at SoundHound AI’s patent filing activity and key business activities from 2015 to 2023 reveals a strategy of balancing innovation with product rollouts and partnerships.
In 2015, SoundHound introduced its core Speech-to-Meaning®, and Deep Meaning Understanding® technologies. From 2017 to 2019, the company focused on integrating voice AI into automobiles through partnerships with Hyundai, Mercedes-Benz, and Honda. These collaborations revolved around the development of intelligent voice assistant systems for these brands through its Houndify platform, which likely prompted the increased patent filing activity that peaked in 2019.
Patent activity declined after 2019 as the company shifted focus to applying its existing voice AI capabilities across new sectors. By then, SoundHound AI had already built a solid patent portfolio around voice recognition, processing, and response. From 2020 onward, SoundHound AI expanded its partnerships into areas like television (VIZIO) and social media (Snap Inc.), by integrating its Houndify platform into these companies’ services.
The slight increase in filings by 2022 coincided with SoundHound AI’s public market debut, which allowed them to raise capital and demonstrate their commitment to innovation. Research and development efforts were demonstrated to show the company’s approach towards innovation and potentially appeal to more investors. This focus on R&D is reflected in their Q1 2022 financial results wherein this category had the highest budget allocation of $14.4 M in 2021 and $16.7M in 2022.
SoundHound AI: Top Jurisdictions
While SoundHound AI supports 25 languages, its patent applications are heavily concentrated in the United States. In addition to being headquartered in California, a number of SoundHound AI’s key partnerships were with U.S.-based companies. A factor that could have also greatly influenced their patent’s jurisdiction was the language- and culture-specific nature of voice AI products. Before expanding globally, SoundHound AI must first develop baseline technologies in one language, then adapt them to the linguistic nuances of other markets.
Nevertheless, SoundHound AI has made strategic international expansions. In 2020, they partnered with Honda Motor Company to integrate the Houndify platform into select Honda electric cars and Honda Jazz models in Europe and Japan. In 2022, they also partnered with Dongfeng Peugeot Citroën Automobiles (DPCA) in China, allowing SoundHound AI to deliver its voice-powered music and navigation capabilities to the Chinese market.
SoundHound AI: Top Law Firms
SoundHound AI’s choice of legal representation for countries outside of the United States has a distinct pattern. For Japan, China, and Korea, they only worked with a single law firm per jurisdiction. Specifically, the 33 patents and patent applications from Japan were all handled by Fukami Patent Office PC. This was also the case for East Intellectual Property in China and FirstLaw PC in Korea. In the United States, SoundHound AI’s filings were distributed across different firms such as Platinum Intellectual Property LLP, Dana Legal Services, and Vierra Magen Marcus LLP (now merged with Pearl Cohen). Notably, the counsel of Vierra Magen Marcus LLP are now members of Pearl Cohen Zedek Latzer Baratz LLP.
SoundHound AI: Top Technology Areas
SoundHound AI’s top technologies are centered on digital data processing (G06F) and speech analysis (G10L). This shows their efforts to improve their core voice AI technologies, specifically as their products rely on the efficiency and accuracy of their voice recognition and processing capabilities.
Moreover, computing arrangements (G06N) and technology systems for commercial purposes (G06Q) were also among their top technology areas, which aligns with their business strategy of integrating their voice AI technology into a variety of commercial settings. This is further supported by the presence of filings related to vehicle technologies (B60R), which is connected with their automotive partnerships.
They also have a number of patents that fall under image recognition (G06V), graphical data presentation (G06K), and image data processing (G06T) which suggests that aside from voice-related innovations, SoundHound AI is also actively developing the visual aspects of its technologies.
Patents behind SoundHound AI’s in-car voice commerce ecosystem
The “commercial” aspect of SoundHound AI’s in-car voice platform makes it distinct from traditional voice assistants. This feature is enabled by various patents for systems that are designed for smoother speech assistance within vehicles.
Voice and gesture controlled interface within a vehicle
Voice control is an optimal form of assistance for drivers given the visual and manual limitations they face while on the road. However, a vehicle’s noisy and high-motion environment poses issues in terms of the accuracy of voice recognition systems.
U.S. Pat. App. No. 2022/0139393 presents an interface system that utilizes audio and visual inputs to better understand a driver’s commands. The visual data gathered by the interface includes facial expressions and gestures, which gives further context to the spoken request of the driver.
The system includes a camera that will capture images of the driver and a microphone for picking up speech. As shown in Fig. 7, the visual and audio inputs will be processed by their respective extractors, and then sent to a linguistic model. This model uses machine learning to process the inputs and enhance the interpretation of the spoken command. The result would be a refined command interpretation that can be turned into an actionable instruction, such as a car system operation or voice feedback.
The patent application titled “Driver interface with voice and gesture control” was filed on December 10, 2021 and was published on May 5, 2022. The inventors are Zili Li and Cristina Vasconcelos. Legal representation was provided by Pearl Cohen Zedel Latzer Baratz LLP.
Optimized voice-based menu ordering
Voice assistance systems can be unreliable in terms of context-specific tasks. Computer-based systems do not always have sufficient context of what is being said by a user, especially in situations that involve specific information, such as ordering from a certain menu. This leads to unsatisfactory responses, which is disadvantageous for businesses that use such kind of voice assistance technology.
Alt text: Block diagram depicting a system that generates an order from a catalog based on speech input
U.S. Patent No. 12,124,804 presents a system that optimizes voice-based menu ordering. A domain-specific catalog, such as a restaurant menu, is first fed into the system. The attributes of the menu items will also be inputted. A “specialist grammar” tailored from this input will then be generated. The process will then follow as illustrated in Fig. 2.
When a user speaks their order, it will be recognized, transcribed, then sent into the “intent engine”. This engine functions to determine the user’s intent based on the combination of the user’s words and the specialist grammar generated from the menu. Aside from matching the spoken phrases with menu items, the intent engine also considers modifiers such as “no” or “with” to understand customized orders (e.g. “no pickles” or “with extra cheese”). Finally, the order generator will take the intent information to format the final order and send it to the restaurant’s point-of-sale system for order processing.
The patent titled “Ordering from a menu using natural language” was filed on April 8, 2022 and was granted on October 22, 2024. The inventors are Joe Kyaw Soe Aung, Vincent Garcia, and Junru Ren. Legal representation was provided by YPS – SoundHound Matters, with attorneys Brian Marcus and Bruce Young.
Engine for upselling in restaurant orders
Traditional voice ordering systems are limited in terms of their capability to understand various natural language expressions and suggest upsells. These systems rely on fixed phrases or item names, making them less adaptive in comparison to how food ordering typically occurs.
U.S. Pat. App. No. 2022/0165272 introduces a recommendation engine that is powered by machine learning (ML). The engine is trained using natural speech inputs collected from customers. It identifies frequently used but unrecognized words, such as slang or indirect phrases. For example, users may say “something healthy” when ordering, and the system may not initially correspond it to a specific menu item.
This phrase will then be flagged for review by a restaurant owner who can map these expressions to specific menu items. In the example, “something healthy” may be mapped to a vegetable salad. Once mapped, the system can then recommend the item when a new user uses a similar language. This will enable the engine to suggest menu items that are statistically likely to be wanted based on the user’s phrases.
Since the engine is powered by ML, the model improves as more inputs are received, making the recommendations more refined over time. In the context of SoundHound AI’ in-car voice commerce system, this feature makes the platform more appealing to restaurants as business partners.
The patent application titled “Recommendation engine for upselling in restaurant orders” was filed on February 8, 2022 and was published on May 26, 2022. The inventors are Kamyar Mohajer and Robert Macrae. Legal representation was provided by Pearl Cohen Zedel Latzer Baratz LLP, with attorneys Larry Vierra, Brian Marcus, and Burt Magen et al. listen on the application.
The road ahead for SoundHound AI’s in-car voice commerce platform
SoundHound AI’s in-car voice commerce system is a part of the broader technological shift towards making commerce more intuitive and seamless. Their patenting activities, such as the refinement of their core technologies to make voice AI more intelligent and efficient, are coordinated with their business strategies, such as partnerships and acquisitions.
Aside from food ordering, SoundHound AI has also announced future applications for this commerce platform, such as car parking, ticket buying, and healthcare. Altogether, these reflect the capability of voice AI to make transactions more personalized and embedded in our daily routines.
