Item 1. Business
Item 1. Business.
Overview
Cerence builds AI powered virtual assistants for the mobility/transportation market. Our primary target is the automobile market, but our solutions can apply to all forms of transportation including but not limited to two-wheel vehicles, planes, tractors, cruise ships and elevators. Our solutions power natural conversational and intuitive interactions between vehicles, drivers and passengers, and the broader digital world. We are a premier provider of AI-powered assistants and innovations for connected and autonomous vehicles, including one of the world’s most popular software platforms for building automotive virtual assistants, such as “ Hey BMW ” and “ Ni hao Banma ”. Our customers include all major automobile original equipment manufacturers, or OEMs, or their tier 1 suppliers worldwide, including BMW, Daimler, FCA Group, Ford, Geely, GM, Renault-Nissan, SAIC, Toyota, Volkswagen Group, Aptiv, Bosch, Continental, DENSO TEN, NIO, XPeng and Harman. We deliver our solutions on a white-label basis, enabling our customers to deliver customized virtual assistants with unique, branded personalities and ultimately strengthening the bond between their brands and end users. Our vision is to enable a more enjoyable, safer journey for everyone.
Our platform utilizes industry-leading speech recognition, natural language understanding, speech signal enhancement, text-to-speech, and acoustic modeling technology to provide a conversational AI-based solution. Virtual assistants built with our platform can enable a wide variety of modes of human-vehicle interaction, including speech, touch, handwriting, gaze tracking and gesture recognition, and can support the integration of third-party virtual assistants into the in-vehicle experience.
Our software platform is a market leader for building integrated, branded and differentiated virtual assistants for automobiles. As a unified platform and common interface for automotive cognitive assistance, our software platform provides OEMs and suppliers with an important control point with respect to the mobility experience and their brand value. Our platform is fully customizable and designed to support our customers in creating their own ecosystem in the automobile and transforming the vehicle into a hub for numerous connected devices and services. Virtual assistants built with our software platform can address user requests across a wide variety of categories, such as navigation, control, media, communication and tools. Our software platform is comprised of edge computing and cloud-connected software components and a software framework linking these components together under a common programming interface. We implement our software platform for our customers through our professional services organization, which works with OEMs and suppliers to optimize our software for the requirements, configurations and acoustic characteristics of specific vehicle models.
The market for automotive cognitive assistance is rapidly expanding. The proliferation of smartphones and smart speakers has encouraged consumers to rely on a growing number of virtual assistants and special-purpose bots for various tasks such as controlling entertainment systems and checking the news. Automobile drivers and passengers increasingly expect hands-free access to virtual assistants as part of the mobility experience, with common use cases in a variety of categories including mobility domains such as navigation, voice-activated texts, and telephone communication, automobile domains, such as automobile user guides, and ignition on-off, and generic domains, such as entertainment. To meet the increasing demand for automotive cognitive assistance and to offer differentiated mobility experiences, OEMs and suppliers are building proprietary virtual assistants into an increasing proportion of their vehicles. We believe that this trend will continue and that consumer appetite for automotive cognitive assistance will grow further as vehicles become more autonomous and drivers pursue new forms of human-vehicle engagement previously not feasible during vehicle operation.
We generate revenue primarily by selling software licenses and cloud-connected services. In addition, we generate professional services revenue from our work with OEMs and suppliers during the design, development and deployment phases of the vehicle model lifecycle and through maintenance and enhancement projects. Through our over 20 years in the automotive industry, we have developed longstanding industry relationships and benefit from incumbency. We have existing relationships with all major OEMs or their tier 1 suppliers, and while our customer contracts vary, they generally represent multi-year engagements, giving us visibility into future revenue. We have master agreements or similar commercial arrangements in place with many of our customers, supporting customer retention over the long term.
As of September 30, 2021, we had fixed backlog of $336.4 million, which includes $276.7 million of estimated future revenue related to remaining performance obligations and $59.7 million of contractual commitments which have not yet been invoiced. As of September 30, 2021, we had variable backlog of $1.7 billion, which includes estimated future revenue from variable forecasted royalties related to our embedded and connected businesses. Our estimation of forecasted royalties is based on our royalty rates for embedded and connected technologies from expected car shipments under our existing contracts over the term of the programs. Anticipated shipments are based on historical shipping experience and current customer projections that management believes are reasonable as of the date of this Form 10-K. Both our embedded and connected technologies are priced and sold on a per-vehicle or device basis, where we receive a single fee for either or both the embedded license and the connected service term. However, our fixed and variable backlog may not be indicative of our actual future revenue. The revenue we actually recognize is subject to several factors, including the number and timing of vehicles our customers ship, potential terminations or changes in scope of customer contracts, and currency fluctuations. As of September 30, 2021, we estimate our total backlog to be $2.0 billion, including $336.4 million of fixed backlog and $1.7 billion of variable backlog.
2
Our solutions have been installed in more than 400 million automobiles to date, including over 40 million new vehicles in fiscal 20 2 1 alone. Based on royalty reports provided by our customers and third-party reports of total vehicle production worldwide, we estimate that approximately 5 3 % of all shipped cars during the fiscal year ended September 30, 20 2 1 included Cerence technologies . Cerence hybrid solutions shipped on approximately 8 . 9 million vehicles during the fiscal year ended September 30, 20 2 1 . In aggregate, over 6 5 automobile brands worldwide use our solutions, covering over 70 languages and dialects, including English, German, Spanish, French, Mandarin, Cantonese and Shanghainese.
In fiscal year 2021, we generated revenue of $387.2 million, an increase of 17.0% compared to $331.0 million for the fiscal year ended September 30, 2020. We recorded net income of $45.9 million for the fiscal year ended September 30, 2021, an increase of 350.8% compared to a net loss of $18.3 million recorded for the fiscal year ended September 30. 2020. For fiscal year 2019, our business was wholly-owned by Nuance Communications, Inc., (“Nuance”), and our results for that fiscal year may not reflect what our results would have been had we been an independent, publicly traded company during fiscal year 2019. In addition, the financial information included herein may not necessarily reflect our results of operations in the future.
History and Corporate Information
On October 1, 2019 (“Distribution Date”), Nuance, a leading provider of speech and language solutions for businesses and consumers around the world, completed the legal and structural separation and distribution to its stockholders of all of the outstanding shares of our common stock, and its consolidated subsidiaries, in a tax free spin-off (“Spin-Off”). The distribution was made in the amount of one share of our common stock for every eight shares of Nuance common stock (“Distribution”) owned by Nuance’s stockholders as of 5:00 p.m. Eastern Time on September 17, 2019, the record date of the Distribution.
In connection with the Distribution, on September 30, 2019, we filed an Amended and Restated Certificate of Incorporation, or the Charter, with the Secretary of State of the State of Delaware, which became effective on October 1, 2019. Our Amended and Restated By-laws also became effective on October 1, 2019. On October 2, 2019, our common stock began regular-way trading on the Nasdaq Global Select Market under the ticker symbol CRNC.
Our principal executive offices are located at 1 Burlington Woods Drive, Suite 301A, Burlington, Massachusetts 01803 and our telephone number at that address is (857) 362-7300. Our website is www.cerence.com . We are not including the information contained in our website as part of, or incorporating it by reference into, this Form 10-K. We make available free of charge through our website our annual reports on Form 10-K, quarterly reports on Form 10-Q, current reports on Form 8-K and amendments to these reports, as soon as reasonably practicable after we electronically file these materials with, or otherwise furnish them to, the Securities and Exchange Commission, or the SEC. The SEC maintains a website ( www.sec.gov ) that contains reports, proxy and information statements, and other information regarding issuers that file electronically with the SEC.
3
Our Capabilities
Our mission is to empower the transportation ecosystem with digital platform solutions for connected and autonomous vehicles. We deliver automotive cognitive assistance solutions that are conversational and intuitive and that enable OEMs to strengthen the emotional connection with their end users through a distinct, consistent, branded experience. We continue to extend these solutions to two-wheel vehicles and tractors and other transportation means. Our principal offering is our software platform, which our customers use to build virtual assistants that can communicate, find information and take action across an expanding variety of categories, including navigation, control, media, communication, information and tools. Our software, developed in deep partnership with the automotive industry, improves the mobility experience for drivers and passengers all over the world.
User engagement with virtual assistants built with our software platform typically begins with a voice request. Upon receiving such an input, our software platform determines what the user has said, infers user intent, and maps the request to the most applicable category and domain. Depending on the applicable domain, our software platform determines whether to respond directly or access an external data source or third-party virtual assistant, in all cases resulting in a response including spoken words or taking action. Depending on the complexity of the request and other factors, engagement may consist of multiple rapid voice interactions with the user and may combine assistance in multiple domains.
Our software platform offers a hybrid architecture combining edge software components, which are embedded in a vehicle’s head unit and integrated with onboard systems, with cloud-connected components, which access data and content on external networks and support over-the-air updates. This hybrid architecture enables our software platform to combine the performance, reliability, efficiency, security and tight vehicular integration of embedded software with the flexibility that cloud connectivity provides. Response frameworks can generally be customized such that requests are processed first at the edge, controlling cloud transmission costs, or in parallel at the edge and in the cloud, to achieve higher confidence responses with low latency. Through edge computing capabilities, the platform is able to provide certain features, such as wake up words, while avoiding privacy and latency issues associated with always-listening cloud-connected technologies. Our software platform includes a common programming framework including toolkits and applications for its edge and cloud-connected components, and our customers can choose the software components that are necessary to power the experiences that they want to build and offer.
4
Cerence Platform Framework - Hybrid Architecture
We deliver our software platform through our professional services organization, which works with OEMs and suppliers to tailor it to the desired requirements, configurations and acoustic characteristics of specific vehicle models. For an initial implementation, our professional services engagements typically begin with the porting of our key technologies to the customer’s specific hardware and software platforms and the development of specific dialogues and grammar libraries. Our professional services teams also work with OEMs on acoustic optimization of a system and application of our audio signal processing technologies. Following an initial implementation, our professional services organization may continue to provide services over the course of a head unit program and vehicle model lifecycle through maintenance and enhancement engagements.
Edge Software Components
Our software platform’s edge software components are installed on a vehicle’s head unit and can operate without access to external networks and information. We tailor our edge software components to a customer’s desired use cases and a vehicle model’s unique systems, sensors and data interfaces.
Capabilities of our edge software components include automatic speech recognition, natural language understanding, noise cancellation, driver and passenger voice isolation, voice biometrics, wake-up word and text-to-speech synthesis, as well as certain non-speech technologies such as gaze, gesture and touch input. Our software can support more than 70 languages and dialects. Edge deployment suits these technologies as it provides the following functionality and benefits:
•
Performance. Processing at the edge is often necessary to meet the low latency requirements of natural conversation.
•
Vehicle Systems Integration. Vehicle applications, sensors, and data interfaces can be integrated deeply with embedded systems.
•
Availability. Edge-located systems are available regardless of cellular coverage and network connectivity.
•
Reduced cost. Processing at the edge reduces or eliminates cellular data transmission costs.
•
Privacy. Users’ utterances and system outputs processed at the edge remain onboard and can immediately be purged.
Certain forms of assistant speech invocation can only be implemented using edge software. The use of wake-up words like “ Hey BMW ” and “ Ni hao Banma ” require constant listening and signal processing to identify instances when a virtual assistant should activate and respond. The same requirements apply to our JustTalk technology, which constantly listens to spoken conversation, determines speaker intent, and invokes assistance appropriately without requiring a specific invocation phase. The alternative of sending a constant stream of audio from the car interior to the cloud for processing would require enormous amounts of bandwidth and potentially create privacy concerns.
We typically sell our edge software components under a traditional per unit perpetual software license model, in which a per unit fee is charged for each software instance installed on an automotive head unit. Our customers generally provide estimates of the units to be shipped for a particular program, and we review third-party market studies and work with our customers to refine and
5
understand these projections. While these projections provide us with some reasonable visibility into future revenue, the number of units to be shipped for a particular program is not committed upfront.
Cloud-Connected Components
Our software platform’s cloud-connected components are comprised of certain speech and natural language understanding related technologies, AI-enabled personalization and context-based response frameworks, and content integration platforms. Our cloud-connected speech-related technologies perform many of the same tasks as our speech-related edge components while offering enhanced functionality through increased computational power and access to external content. Cloud-connected components also support the replication of personalized settings such as voice profiles and preferences across multiple vehicles.
We offer cloud-connected components in the form of a connected service to the vehicle end user. Initial subscriptions typically have multi-year terms from the time of a vehicle’s sale and are paid in advance by the OEM or supplier. Renewal options vary and are managed by our customers on behalf of vehicle end users.
Virtual Assistant Coexistence
The wide variety of use cases encompassed by automotive cognitive assistance, in the context of evolving consumer preferences, necessitates the coexistence of multiple virtual assistants within the in-vehicle environment. For example, many vehicle-related categories such as navigation and control can best be addressed by a tightly integrated, vehicle-model-specific virtual assistant. At the same time, drivers and passengers often prefer to use familiar Internet-based virtual assistants for more general domains such as entertainment.
To enable drivers and passengers to extend their digital life from outside the vehicle to inside the vehicle, our software platform can support the integration of third-party virtual assistants, providing a uniform interface for virtual assistant engagement. We have invested in our platform to develop the technology and capabilities necessary to integrate third party virtual assistants with vehicles’ systems.
To make integration as seamless as possible, we have built cognitive arbitration technology that is capable of inferring user intent, determining which within a set of virtual assistants would be best suited to address a request, and sending the request to the selected assistant thus enabling users to extend their digital life into the automobile. Depending on a system’s configuration and the virtual assistants to which it is connected, output can be presented back to the user through a vehicle-specific personality or through the virtual assistant’s own interface. Cognitive arbitration represents an important control point with respect to the mobility experience and an important brand differentiation opportunity for OEMs and suppliers. Like the rest of our software platform, cognitive arbitration is a white label product that can be customized and branded.
Along with providing OEMs control over their brand identity, our cognitive arbitration technology is an important element in letting an OEM design the overall driver and passenger experience. This technology allows an OEM to dictate interactions with third-party virtual assistants within the vehicle, strengthening its ability to differentiate and control the overall in-vehicle experience.
Professional Services
We have a large professional services team that works with our customers in the design, development and deployment phases of a vehicle head unit program and vehicle model lifecycle, as well as in maintenance and enhancement engagements. Our range of capabilities include personalization of grammar and natural language understanding development, localization, language selection and system coverage, navigation speech data generation, system prompt recordings, porting our platform’s framework and our ability to deploy cognitive arbitration technologies, and user experience reviews and studies. Our professional services team is globally distributed to serve our customers in their primary design and production jurisdictions. We typically charge manufacturers for our design and consulting work, which are primarily project-based, in line with customary non-recurring engineering industry practices.
6
Our Competitive Strengths
Our key competitive strengths include:
•
Industry-leading speech-related technology . Our research shows that consumers see speech as an increasingly attractive medium for human-vehicle interaction. Nevertheless, they are often frustrated with speech recognition solutions that misunderstand spoken language or require users to speak rigid, pre-defined commands associated with a limited set of functions. Developing conversation-based automotive virtual assistants that users will perceive as natural is challenging as a matter of artificial intelligence technology, acoustic engineering and user interface design. We believe our software platform, as tailored for a specific vehicle model by our professional services organization, represents one of the most technologically advanced and highest-performing human-vehicle speech interaction systems available today. In tests performed by our customers to assess correct recognition of words, sentences, and domains, our solutions have achieved some of the highest marks relative to competitors and our offerings are backed by our portfolio of patents and associated rights.
•
Hybrid edge-cloud system architecture. Our software platform’s hybrid architecture combines the performance, reliability and tight integration that only edge software can provide with the flexibility of cloud connectivity. Cloud-reliant solutions with which our software platform competes cannot match edge software’s low latency, its bandwidth efficiency or its availability in the absence of network connectivity. Moreover, emerging speech invocation paradigms such as wake up words and situationally aware invocation are most effectively implemented using edge technology.
•
Bespoke vehicle integration and acoustic tuning . Cognitive assistance for categories such as navigation, entertainment and control requires tight integration with onboard vehicle components, which vary widely among vehicle models. Separately, speech interaction systems can be significantly hampered by the noisy environment of a vehicle cabin and must be tuned for particular acoustics and audio system components. To achieve the tight vehicle integration necessary to address these concerns, our professional services team works closely with OEMs and suppliers to customize our offerings for the particular characteristics of specific vehicle models. Our expertise in acoustics enables us to implement systems that can isolate the voices of individual speakers and support simultaneous virtual assistance for speakers in multiple zones, representing a key point of differentiation.
•
Interoperability with third-party Internet-based virtual assistants. Virtual assistants from large technology companies have become popular with consumers. We believe that consumers want to extend the use of these assistants while traveling in their vehicles and that a comprehensive automotive cognitive assistance system requires the coexistence of multiple virtual assistants. To accommodate their end user preferences while still providing a unique and brand-specific experience, OEMs seek to offer a common in-vehicle interface with seamless integration across various virtual assistants. To this end, our software platform can support the coexistence of multiple third-party virtual assistants and provide a uniform interface for virtual assistant engagement. Our market-leading position, our focus on the automotive market and the large size of our installed base create incentives for third party virtual assistant providers to work with us and support this integration.
•
Independence from large technology companies and automobile industry players. As vehicles become more autonomous, mobility experiences are being increasingly defined by in-cabin features and alternative forms of human-vehicle engagement. Branded, differentiated automotive cognitive assistance is thus increasingly important to OEMs’ brand value. As a neutral, independent, white-label software platform vendor, we empower our customers to build branded and differentiated experiences and retain ownership of, or rights to, their system design and data. The virtual assistant coexistence enabled by our cognitive arbitration functionality is designed to allow our customers to provide access to third-party virtual assistants without ceding overall control of the cognitive assistance experience.
•
OEM alignment . The design and development of the head unit within the vehicle ecosystem is a complex process requiring tight integration of the software and hardware components used in and with the vehicle. We believe our demonstrated long-standing capabilities in working closely with OEMs, understanding their needs, product roadmaps and global go-to-market strategies enables us to innovate our technologies to meet an OEM’s specifications. Furthermore, our working relationships with OEMs uniquely allow us to market and sell our solutions on both a local and global basis in accordance with an OEM’s particular requirements.
•
Broad language coverage. Our software platform supports over 70 languages and dialects, far more than any of our competitors. As a result of our broad language support, our customers are already delivering cognitive assistance based on our software platform across the Americas, Europe and Asia, including China, the U.S. and all other large automotive markets. Our language support also enables multi-lingual capabilities for domains such as music selection, point-of-interest selection, and cross-border navigation among others, representing a critical feature for markets such as Continental Europe in which automobiles may routinely traverse multiple lingual zones. We believe that our portfolio of languages and multi-lingual capabilities represent an important competitive advantage, as the development of capabilities to support a new language is expensive and time-consuming.
7
•
Broad, global network of deep relationships with OEMs and tier 1 suppliers. We have supplied speech recognition systems to OEMs and suppliers for over 20 years, working closely with our customers through our global professional services organization to design and integrate our solutions into their brands. Today, we work with all major OEMs or their tier 1 suppliers worldwide, leveraging the geographic breadth and industry experience of our professional services teams. Our long history in the automotive industry and the global reach and experience of our over 5 0 0 professional services employees across 1 2 countries gives us credibility with OEMs as we seek new business with OEMs, either directly or through their tier 1 suppliers. We believe that OEMs who sell globally will value our experience in servicing and deploying solutions on a global basis. We often have master agreements or similar commercial arrangements with our customers. These master agreements help us retain customer relationships over the long term.
Our Growth Strategies
We believe our growth opportunity has three key facets: continued investment in expending our core technology, development of new applications that extend our core technology into innovative applications, and expansion of our target market beyond automobiles. Successful execution of these key objectives could lead to the greater penetration of our offerings and key enabling technologies throughout our target markets, resulting in an increase in the revenue we are able to capture per vehicle and expansion of our market share relative to competitors.
Our primary strategies for pursuing our growth include the following:
•
Maintain and extend product leadership . We intend to continue investing in developing our core product functionality and expanding the breadth of categories and domains our software platform is able to address, particularly with a view toward maintaining our market share in edge software components and growing our share in cloud-connected software functionalities. Our existing relationship with, and our proximity in the design process to, OEMs provides us with insight into the needs of the end-users and roadmaps for innovation. For instance, this insight has helped us identify and advance our technologies for autonomous driving systems, which technologies have been incorporated in solutions currently under development. Additionally, we intend to continue to invest in customizing and supporting our solutions for specific individual automobile vehicle models, resulting in tight integration of our solutions. We believe that increasing complexity of our edge software components, including with respect to multi-modal interaction, and growth in our cloud-connected product areas, including the enabling of third-party services, will enable us to increase the revenue per vehicle that we are
8
able to generate. Additionally, these investments will help maintain our position with existing customers through new vehicle models and enable us to grow with the overall market for automotive cognitive assistance.
•
Continue to invest in interoperability with third-party virtual assistants . We believe that the growing popularity of third-party virtual assistants is creating increasing demand for access to these assistants as part of the mobility experience. We also believe that complete automotive cognitive assistance requires the coexistence of multiple virtual assistants. We intend to continue to invest to develop our software platform’s interoperability with third-party virtual assistants and its cognitive arbitration capabilities to maintain its position as a neutral automotive cognitive assistance platform. We believe a neutral automotive cognitive assistance platform will increasingly be valued by OEMs that prioritize maintaining their unique and branded in-car experience and the ability to control the mobility experience overall.
•
Deliver new functionality to existing installed base. Our solutions have been installed in more than 400 million vehicles to date. Our large installed base represents an opportunity to deliver new features and software. Depending on system capabilities, we are able to deliver updated functionality to our users in the form of embedded software upgrades performed by dealers and over-the-air updates delivered from the cloud.
•
Develop products that leverage our expertise in new applications. We have developed new products that leverage our expertise in voice-AI into new applications that will be distinct from our Edge or Cloud-connected product offerings. These new applications are expected to generate revenue using either a subscription or transaction-based model extending the company’s market opportunity into new areas. New applications developed include Cerence Tour Guide, Cerence Pay and Car Life. Cerence Tour Guide is an AI-powered application for automotive assistants that brings guided tour content directly into the car via an ecosystem of partners. Cerence Pay offers a secure, contactless payment experience for drivers via voice and facial biometrics. Cerence Car Life is a suite of AI-powered, software-as-a-service (SaaS) offerings that provides drivers with up-to-date information about their cars via a companion application, voice output from the automotive assistant, and imagery displayed on the car’s infotainment system.
•
Expand into adjacent transportation markets. Today, we primarily target the automobile market. However, our products and technology also have application to other modes of transportation. Any type of vehicle that moves people are potential applications for our technology. We have integrated our technologies and solutions within the two-wheel vehicle market and have explored opportunities in the cruise line, public transit, fleet, and elevator markets. In total, we believe these adjacent markets represent an important growth opportunity.
Competition
The automobile cognitive assistance market is competitive. Today, we face two primary sets of competitors:
•
Large technology companies . Many large technology companies, including Amazon, Apple, Google, Microsoft, Alibaba, Baidu and Tencent, offer Internet-based virtual assistants. Given the popularity in general of these virtual assistants, we believe that automobile drivers and riders increasingly desire the ability to use them as part of the mobility experience. To meet this demand, some of these companies have invested in technologies, such as Apple CarPlay, to make their virtual assistants more accessible within vehicle cabins.
While these third-party virtual assistants directly compete with some of the functionality we provide as part of our software platform, they also increase the need for our software platform in two ways. First, given the fragmented and competitive nature of the virtual assistant market, it is important for OEMs and suppliers to enable their passengers to utilize a variety of virtual assistants. Our software platform’s cognitive arbitration functionality can, dependent on appropriate third-party agreements, enable OEMs and suppliers to provide access to multiple third-party virtual assistants through a consistent, branded interface. Second, the noisy environment of a vehicle cabin presents significant speech processing challenges for smartphone-based third-party virtual assistants that are not designed for a specific vehicle model. Our software platform integrates with third-party virtual assistants and improves their functionality by improving the quality of speech input.
•
Small, focused competitors . We compete for business directly with certain companies focused on voice-based virtual assistance, including SoundHound in the U.S., iFlyTek in China, and other regional and technology-focused competitors. These companies have had some success selling into our customer base. However, we believe that we have multiple meaningful competitive advantages, including our scale, our globally distributed team, our best-in-class portfolio of compatible languages, and our deep focus on the automotive market. We also believe that our technology, particularly our speech signal enhancement and acoustic tuning, is superior based on benchmarking results against our competitors. We believe we will continue to be able to compete successfully against these competitors as we continue to invest in our offerings.
Our industry has attracted, and may continue to attract, new entrants. Although we find that OEMs often prefer to maintain relationships with suppliers that have a proven record of performance, they also rigorously reevaluate suppliers on the basis of product quality, price, reliability and timeliness of delivery, product design capability, technical expertise and
9
development capability, new product innovation, financial viability, operational flexibility, customer service and overall management.
Technology
Our software platform’s edge and cloud-connected software components are based on a number of proprietary technologies. We customize these technologies for specific vehicle models and continuously update and improve our features and functionality. Our key technologies include but are not limited to the following:
•
Speech Signal Enhancement . A high-quality voice input signal is a precondition to reliable speech recognition and cognitive assistance. However, in a typical vehicle cabin, ambient interior sounds and noise from around the vehicle mix with infotainment system output and conversations between passengers, create a complex soundscape that can obscure virtual assistant requests. Audio signal processing technologies are therefore critical to the cognitive assistance experience. We have been developing and combining highly advanced audio signal enhancement technologies for over 20 years, and we tune our software in relation to the placement of microphones in a vehicle to create defined acoustic zones and support the isolation of individual speakers. Our technologies deliver best-in-class speech recognition results, as evidenced by tests performed by our customers to assess correct recognition of words, sentences, and domains, in which our solutions have achieved some of the highest marks relative to competitors.
•
Automatic Speech Recognition . Our speech recognition technology, built using neural networks and specifically designed for automotive applications, is recognized as the automotive industry leader in automatic speech recognition. We support over 70 languages and dialects, representing the largest language portfolios in the speech industry. Key features of our speech recognition technology include free-form conversational interpretation, as opposed to a rigid system of predefined commands, and barge-in capabilities, enabling users to correct and modify their requests in the middle of stating them.
•
Natural Language Understanding . Once speech has been captured and accurately converted into words, natural language understanding technology, or NLU, is necessary to match the request to the appropriate category and domain to interpret the user’s intent. Our NLU system applies artificial intelligence reasoning, including predefined and learned preferences and real-time contextual information, to deliver informative responses consistent with what a user desires. NLU processing is performed by a hybrid of edge and cloud-connected software components to optimize performance, efficiency, reliability and security.
•
Vocalizer: Text-to-Speech and Natural Language Generation . In many cases, the most useful result of a spoken query or command is a spoken response back to the user. To enable cognitive assistants to speak, we offer text-to-speech technology in more than 65 languages and dialects and over 145 distinct voices. We also have developed the technology to read text using human-like inflection and emotion, as well as, offer custom voices for customers who wish to differentiate themselves through an exclusive personality representing their brand.
•
Voice Biometrics . Our software platform includes biometric functionality which can authenticate and personalize the automotive experience by recognizing users based on their voice and automatically load individual preferences and other automotive settings.
•
Push-to-Talk, Wake-Up Words and Just Talk . Through our software platform, we are capable of offering three methods for invoking the virtual assistant, which can be implemented alone or in combination:
•
Push-to-Talk functionality, most commonly implemented as a button on the steering wheel or center console.
•
Wake-Up Word functionality, involving a spoken keyword or phrase, such as “ Hey BMW .”
•
Just Talk . Our active listening technology, filters out background noise and irrelevant conversation until it hears a keyword, phrase, or command that it understands as related to an applicable domain and which is intended as a virtual assistant request. False triggers are minimized through sophisticated syntax, cadence and intonation analysis performed in real-time and can be further reduced using automobile sensors such as head or body movement trackers.
•
Cognitive Arbitration. Our cognitive arbitration technology can route arbitrary requests to the most appropriate virtual assistant or bot, including third-party virtual assistants.
•
Non-Speech, Multimodal Input. Our technology seeks to mimic conversational human interaction by incorporating input methods beyond speech. Our multimodal capabilities allow vehicle systems to accept multiple forms of input such as voice, gestures, gaze, predictive text and handwriting.
•
Multi-Seat Intelligence . Due to its flexible design, our speech signal enhancement technology can be easily configured for complex multi-zone scenarios with various users and nearly arbitrary microphone configurations. Dedicated processing modes enable efficient and robust multi-user speech recognition in challenging acoustical environments. This allows for
10
passenger interaction in individual zones like sharing music or interacting simultaneously with the car or infotainment systems, where some passengers can enjoy browsing their music by speech, while others can send emails or other work-related activities.
Research and Development
We maintain technical engineering centers in major regions of the world that help develop our software platform and its underlying components and provide our customers with local engineering capabilities and design development.
We employ approximately 900 research and development personnel around the world, including scientists, engineers and technicians. Our total research and development expenses were approximately $112.1 million, $88.9 million and $93.1 million for fiscal years 2021, 2020 and 2019, respectively.
We believe that continued investment in research and development will be critical for us to continue to deliver market-leading solutions for automotive cognitive assistance. Accordingly, we intend to continue to invest in our product portfolio and allocate capital and resources to our growth opportunities.
Customers
Our customers include all major OEMs or their tier 1 suppliers worldwide. Our automobile manufacturer customers, commonly referred to as OEMs, include BMW, XPeng, FCA Group, Ford, Daimler, Geely, Renault-Nissan, SAIC, Toyota, Volkswagen Group and many others and represented approximately 46% of our sales in fiscal year 2021. Our largest customer, Toyota, represented approximately 19% of our revenue in fiscal year 2021. Our tier 1 supplier customers, who typically sell automobile components to the OEMs, include Aptiv, Bosch, Continental, DENSO TEN, NIO, Harman and many others and represented approximately 54% of our business in fiscal year 2021.
Our revenue base is geographically diverse. In fiscal 2021, approximately 35%, 37% and 28% of our revenue came from the Americas, Europe and Asia, respectively.
Sales and Marketing and Professional Services
We market our offerings using a high-touch OEM solutions model. We sell directly to our customers, which include OEMs and suppliers and as described above under “Customers”, and for each of our customers we assign a team comprising sales and marketing as well as professional services personnel. Our customer contracts are bespoke and vary widely, but generally represent multi-year agreements providing visibility into future revenue and helping to support retention of customer relationships over the long term.
Our sales and marketing team includes approximately 100 employees. This team includes sales representatives, account managers, sales engineers, product managers and marketing experts. As we sell our offerings to all major OEMs or their tier 1 suppliers today, our sales strategy is primarily focused on leveraging our existing customer relationships. Account managers typically have longstanding relationships with specific customers and are distributed worldwide to provide local customer coverage. We oftentimes utilize customer-specific demo days and proof-of-concepts (“POCs”) in which we showcase our technology and capabilities to OEMs and tier 1 suppliers on an individual basis. These events help maintain our market presence and awareness of our platform’s offerings while also providing opportunities to solicit feedback and input from our customers on our roadmap and future technologies.
Our professional services organization includes approximately 500 employees. These employees work with our customers in the design phase of the vehicle lifecycle to tailor our platform for specific requirements such as branding and also tune the software for the characteristics of a vehicle model. Our professional services team also provides post-design phase services through maintenance engagements, particular with respect to our cloud-connected solutions. The tight integration of our platform into our customers’ design process and their vehicles supports our ability to win future business with those customers. Like our sales representatives, our professional services employees often have longstanding relationships with specific customers and are distributed worldwide to provide local customer coverage.
Human Capital
Summary
As of September 30, 2021, we had approximately 1,700 full-time employees, including approximately 100 in sales and marketing, approximately 200 in administrative functions, approximately 500 in professional services, and approximately 900 in research and development. Approximately 90% of our employees are based outside of the United States. None of our employees in the United States are represented by a labor union, however many of our employees in Europe are represented by workers councils or labor unions. To date, we have experienced no work stoppages and believe that we have a good relationship with our employees.
11
Culture and Work Environment
We’re a group of highly motivated collaborators who share a common passion for creating meaningful change in our industry and shaping the future of mobility. We are committed to attracting and retaining the best and brightest talent and building a culture of transparency, trust, and respect.
We are proactively nurturing our culture by investing in our people, processes and professional development. We understand our people are critical for our continued success and are focused on helping our employees grow at every stage of their career. We have education opportunities and training and development programs that help to enrich the knowledge and talents across the organization. From wellbeing programs and holiday celebrations to our virtual book club and LBGTQ alliance, we’re focused on maintaining our connections regardless of our physical locations.
Compensation, Rewards and Benefits
In addition to competitive base salaries, we provide incentive-based compensation programs to reward performance relative to key metrics. We also provide compensation in the form of restricted stock unit grants as well as a competitive time-off policy. We offer comprehensive benefit options, including retirement savings plans, medical insurance, dental insurance, vision insurance, life and disability insurance, health savings accounts, flexible spending accounts, and an employee stock purchase plan, among others.
Diversity and Inclusion
We are a global team that seeks to build a diverse and inclusive workplace built upon the different perspectives, beliefs, and backgrounds of our people. We embrace what makes us each unique. Strengthening diversity enables us to bring our collective ideas together to make the best decisions for the global community we serve. We have successfully launched affinity groups for Diversity and Inclusion , Women in Technology , and Working Parents , as well as our Book Club . We celebrated important cultural observances such as Black History Month, Women’s History Month , and Pride Month . It’s extremely important that every employee feel welcome and valued as we strive to make our company a great place to work.
Intellectual Property
We own approximately 876 patents and patent applications and other intellectual property. Prior to our Spin-Off from Nuance, we entered into the Intellectual Property Agreement, which provides us with certain non-exclusive rights with respect to patents that will continue to be held by Nuance. While no individual patent or group of patents, taken alone, is considered material to our business, in the aggregate, these patents and rights provide meaningful protection for our products, technologies, and technical innovations.
12
Text extracted from the filing as submitted to EDGAR. Formatting, tables and exhibits are simplified for reading; the original document is authoritative for anything you rely on.