Veritone S-1: Read the Full IPO Prospectus (2017)

Prospectus Summary

PROSPECTUS SUMMARY This summary highlights information contained elsewhere in this prospectus. This summary does not contain all of the information you should consider before investing in our common stock. You should read this entire prospectus carefully, including the sections entitled “Risk Factors” and “Management’s Discussion and Analysis of Financial Condition and Results of Operations” and our consolidated financial statements and the related notes appearing at the end of this prospectus, before making an investment decision.

Unless otherwise indicated herein, references in this prospectus to “Veritone,” the “Company,” “we,” “us” and “our” refer to Veritone, Inc., a Delaware corporation, together with its consolidated subsidiaries. References herein to “fiscal year” refer to our fiscal years, which end on December 31.

Company Overview

Veritone has developed a proprietary artificial intelligence platform that unlocks the power of cognitive computing to transform unstructured audio and video data and analyze it in conjunction with structured data in a seamless, automated manner to generate actionable intelligence. Our cloud-based open platform integrates and orchestrates an ecosystem of best-of-breed cognitive engines to reveal valuable multivariate insights from vast amounts of audio, video and structured data.

Our platform incorporates proprietary technology to manage and integrate a wide variety of AI processes to mimic human cognitive functions such as perception, reasoning, prediction and problem solving in order to quickly, efficiently and cost effectively transform unstructured data into structured data. Our platform stores the results in a searchable, time-correlated database, creating a rich, online, searchable index of audio, video and structured data, enabling users to analyze the information in near real-time to drive business decisions and insights.

Our mission is to bring the power of AI-based computing to enterprises of all sizes through a cloud-based platform that integrates multiple cognitive engine capabilities in a single platform that is licensed to our customers via a Software-as-a-Service (SaaS) model. Our scalable platform is based on an open architecture that enables new cognitive engines to be added quickly and efficiently, resulting in a future proof, scalable and evolving solution that can be easily leveraged in a broad range of industries that capture or use audio and video data, including, without limitation, media, politics, legal, law enforcement, retail, and other vertical markets.

Industry Background

According to Gartner, Inc., 80% of the world’s data is unstructured and cannot be easily and efficiently searched, accessed or utilized. Unstructured data continues to grow at an accelerating rate and represents the vast majority of all data created, with 90% of the world’s data being created in the past two years. International Data Corporation (IDC) predicts that the total amount of digital data created worldwide will grow from 4.4 zettabytes in 2013 to 44 zettabytes in 2020. Enterprises, including businesses and government agencies, recognize the utility of leveraging this rapid proliferation of data and are turning to artificial intelligence and analytics to increase their capabilities and efficiency. However, such enterprises lack the proper tools to collect and analyze the rapidly growing variety, velocity and volume of data in real-time.

According to IDC, the worldwide revenue from the market for cognitive systems and artificial intelligence is expected to grow from approximately $8 billion in 2016 to more than $47 billion in 2020. We believe that our platform and related services address a subset of such overall AI market. AI systems, particularly cognitive systems based upon machine and deep learning, are increasingly being designed to adapt and make sense of the complexity and unpredictability of unstructured data, such as text, audio and video. Deep learning enables computers to discern patterns in large sets of data without being told what to look for. These machine learning algorithms can perform a wide range of functions, such as “reading” text, “seeing” images and “hearing” natural speech, and in turn interpret that information, organize it and offer explanations of what it means, along with the rationale, conclusions and applicable recommendations for such information. These systems can also improve the accuracy of their results with training and use. However, current machine learning algorithms are generally only effective when applied to a single, narrowly defined problem, which limits the amount and type of metadata that can be extracted and the nature of inquiries and analyses that can be performed, and therefore limits the usefulness of any single cognitive engine in real-world applications. For companies looking to harness the power of their unstructured data, this presents significant challenges. Many applications require multiple cognitive processes to be performed on unstructured data, and for the results to be correlated with structured data of the user or third parties. In addition, identifying the best algorithm to perform each task can be very challenging as both the user’s needs and the relative performance of algorithms change over time, and switching algorithms within a closed system can be expensive and time consuming.

Our Solution

Veritone’s innovative AI-based open platform intelligently orchestrates multiple best-of-breed cognitive engines within a single cloud-based solution to process information in volumes that can far exceed human cognitive capabilities. Our proprietary technology enables users to run comprehensive, multivariate queries, correlations and analyses in near real-time using multiple cognitive engines and data sets, integrating and refining the outputs. This enables our users to gain more accurate and in-depth analysis, more efficiently and cost effectively than any single engine solution. Our solution can also ingest and analyze structured data in correlation with the processed data. For example, our media users can use our system to identify instances where advertiser names or logos appear in a broadcast, and then correlate those instances with Nielsen data to measure the number of impressions generated by the ad, web traffic data to estimate the traffic driven by the ad, or their own sales data, to provide actionable intelligence regarding their advertising campaigns.

Our innovative AI ecosystem currently incorporates over 40 cognitive engines of various classes and types, from multiple vendors, including Google, IBM, Microsoft, Nuance, and Hewlett Packard, among others, in addition to several Veritone developed cognitive engines. These engines use advanced algorithms to capture and extract data from the contents of such media files for a variety of cognitive capabilities including, but not limited to, transcription, face detection, face recognition, object recognition, sentiment analysis, language translation, audio/video fingerprinting, geolocation, optical character recognition, metadata extraction, and media format transcoding. Our applications then allow users to perform analytics on such data. By having a number of engines from different providers within the same class and across different classes, our AI system can generate extensive and varied training data to more efficiently learn and thereby quickly improve accuracy and analyses. A key principle of AI-based systems is that the more data that is ingested and processed, the better the accuracy and predictive value that they can achieve.

Our Competitive Strengths

Our key competitive strengths include the following:

Proprietary Conductor Technology. Our Conductor technology optimizes the cognitive engine selection process by periodically evaluating the cost, speed, accuracy and usefulness of each engine to improve results and achieve higher levels of cognition. These Conductors orchestrate and intelligently route media to the most appropriate cognitive engines within chosen engine classes to produce robust intelligence from their media. We have released our Conductor technology for transcription and plan to deploy Conductors for other cognitive engine classes, such as object and logo recognition. Because we maintain proprietary performance data on the relative performance of each cognitive engine in our ecosystem, we believe that, in the near future, we will be able to automatically select, in near real-time, the best engines in each major cognitive class to deploy based upon our automated review of the data being processed, including any metadata therein.

Broad Indexing and Search Capabilities. Utilizing our proprietary temporal elastic database (TED), we have the unique ability to synthesize various disparate data sets in a cohesive, time-based format, allowing users to unearth intelligence from the data that was previously unattainable. One of our applications, Veritone Discovery, allows for immediate access to data through dynamic, multivariate querying of various metadata derived from the original data. Elements include, but are not limited to, cognitive engine outputs (transcripts, translations, recognized objects and faces, etc.), file metadata and user-associated metadata such as content template outputs and freeform tags, as well as structured data such as media viewing data, web traffic data, sales data or crime statistics. These disparate data sets are indexed in TED in a time-correlated manner, allowing them to be synthesized in a cohesive, time-based approach. While storage agnostic, TED currently runs on Elasticsearch, leveraging the Apache Lucerne inverted index architecture.

One Open Platform, Integrated Multi-Cognitive Processing. Our open platform allows third-party AI engines to be easily integrated onto our platform and to participate in our AI marketplace via application programming interfaces (APIs). We currently offer over 40 cognitive engines on our platform across most major categories of AI. Our goal is to create a broad ecosystem offering all major cognitive engine capabilities on a single open and integrated platform. This business model enables our customers to send their data to one cloud-based platform, utilize multiple cognitive processes and correlate the resulting disparate metadata from varying third-party engines, as well as structured data sources, and analyze, capture and share the results without leaving their Veritone user session. As our business expands, we plan to increase the number of supported instances of our platform within public and private cloud environments and across multiple geographies.

Cloud Agnostic. Our platform currently leverages several cloud computing and storage services, including Joyent and Amazon’s Web Services cloud. We have also recently deployed our platform to run on Microsoft Azure’s secure government cloud, which enables us to address the needs of the law enforcement and public safety markets. We believe we will be able to deploy our platform on a variety of cloud infrastructures and within varying geographic regions and countries, which will be necessary for our future international expansion.

Modular and Industry Specific Applications. Our modular platform is designed to enable our customers and developers to rapidly develop applications that can convert customer data into actionable insights relevant for different markets. We have created and demonstrated successful use cases for our media, politics and legal applications, which enable customers to solve difficult problems such as programmatic ad verification, tracking audience measurement correlations, and finding, selecting, and editing long-form media content into short-form media shareable via social media.

Our Growth Strategies

We plan to expand our market presence and business by pursuing the following growth strategies:

Our Target Markets

While AI can be applied to a myriad of industries and functions, we have initially targeted a few key markets where users generate large amounts of unstructured audio and video data, and where we believe that our platform has the potential to unlock the power of that data to significantly improve the efficiency and effectiveness of their operations.

Development of our AI Platform Business

Since our inception, we have generated substantially all of our revenues from our media placement services performed under advertising contracts with our media clients (the Media Agency Business). We are in the early stages of developing our AI platform and SaaS licensing business (the AI Platform Business) and have not generated significant revenue from this business. During the years ended December 31, 2016 and 2015, our net revenues were $8.9 million and $13.9 million, respectively, which consisted of $8.4 million and $13.8 million, respectively, of net revenues from our Media Agency Business and $0.5 million and less than $0.1 million, respectively, from our AI Platform Business. We intend to continue to invest significant resources and capital in developing our AI Platform Business, and therefore we do not expect to achieve profitability for the foreseeable future. Our objective is to manage our resources so that we can continue to develop a successful AI Platform Business that generates significant increases in net revenues and establishes us as the leader in the cloud-based artificial intelligence market, while at the same time maintaining and expanding our Media Agency Business by acquiring new customers and adding additional media placement contracts.

Summary Risk Factors

Our business is subject to numerous risks and uncertainties of which you should be aware before you decide to invest in our common stock. These risks may prevent us from achieving our business objectives, and may materially and adversely affect our business, financial condition, results of operations and prospects. These risks are discussed in greater detail in the section entitled “Risk Factors” immediately following this prospectus summary. Some of these risks are:

Implications of Being an Emerging Growth Company

We are an “emerging growth company,” as defined in the Jumpstart Our Business Startups Act of 2012 (the JOBS Act), and are eligible to take advantage of certain exemptions from various reporting requirements and may be relieved of other significant requirements that are otherwise generally applicable to other public companies that are not emerging growth companies. These exemptions include, but are not limited to:

We may take advantage of these exemptions for up to five years or such earlier time that we are no longer an emerging growth company. We would cease to be an emerging growth company if we have more than $1.0 billion in annual revenue, we are deemed to be a large accelerated filer under rules of the Securities and Exchange Commission (SEC) or we issue more than $1.0 billion of non-convertible debt over a three-year period. We have taken advantage of certain reduced reporting burdens in this registration statement, of which this prospectus is a part, and may elect to take advantage of some or all of the reduced reporting requirements in our future filings. As a result, the information contained herein and that we provide to our stockholders in the future may be different than the information you receive from other public companies in which you hold stock.

In addition, Section 107 of the JOBS Act provides that an emerging growth company can take advantage of the extended transition period provided in Section 7(a)(2)(B) of the Securities Act of 1933, as amended (the Securities Act) for complying with new or revised accounting standards. In other words, an emerging growth company can delay the adoption of certain accounting standards until those standards would otherwise apply to private companies. We intend to take advantage of such extended transition period. Section 107 of the JOBS Act provides that we can elect to opt out of the extended transition period at any time, which election is irrevocable.

Channels for Disclosure of Information

Investors, the media and others should note that, following the completion of this offering, we intend to announce material information to the public through filings with the SEC, the investor relations page on our website, press releases and public conference calls and webcasts.