TechBytes with Chris Ricciuti, VP of Product, Financial Services at Veritone

How would you best describe your journey through technology?

My journey through technology is based on diverse experiences. I got my start in technology shortly after electronic trading began to permeate the financial services industry, focusing first on helping to create the plumbing that connects trading floors to exchanges and ultimately working on algorithmic trading systems in their early days. Eventually, I caught the startup bug and launched a company that built regulatory compliance software for financial services firms, which led me to a career at established Silicon Valley software companies. By that point, I started to think about how nascent advances in AI (plus increased data availability and computing power) could be applied to improve the business processes that I had exposure to earlier in my career.

What do you find more interesting, founding companies or working for them? What are the key challenges in both?

To me, both founding companies and working for companies are interesting in their own ways.

Founding a company is interesting in the sense that it’s extremely exciting to work on challenging problems and iterate quickly, as you struggle to keep the lights on. This is especially true in cases where the startup is genuinely innovating and blazing a path into uncharted territory. Startups certainly demand a wide range of skills and require the ultimate dedication and tenacity to succeed. As evidenced by the estimated failure rate of startups – 50% fold within five years and more than 70% fail after 10 years – this is indeed a challenge. That said, the journey – especially when it leads to success – can be extremely rewarding.

When working for established companies, there are theoretically more resources available at your disposal. In terms of human resources, the diversity of skill-sets and perspectives of extremely creative minds, an interesting dynamic exists within more established organizations that has the potential to foster innovation. But in larger organizations, it’s often difficult to move the needle as quickly as in startups, since a wide range of teams need to be consulted on each initiative – each with its own set of competing priorities. Often, politics and the need to earn buy-in from various constituents plays too large a role, which can slow innovation and progress.

What are some of the major roadblocks in integrating AI into the legal and compliance industries?

AI is positioned to disrupt both eDiscovery (within the legal industry) and surveillance (within the compliance sector) processes and workflows by making it easier, quicker and more cost-effective to find the needle in the haystack. Because of this, AI has the potential to move humans from in the loop (where work is quite manual), to on the loop (where AI does the heavy lifting while humans monitor the AI), to out of the loop (where humans are removed from the process entirely). AI’s biggest roadblock might be its own potential, as it can be perceived as a threat, rather than embraced as an opportunity.

Of course, there are day-to-day roadblocks associated with employing various cognitive models, including access to clean data, organizational structures (especially for non-digital native companies), training and adoption, but these can be overcome.

How do you select your markets (EMEA, APAC, North America, etc.) to monetize Veritone’s legal and compliance products? Do you have to build separate products depending on the legal and compliance framework of a particular region?

In our legal and compliance vertical, we select markets based upon the prospect of future demand that we see in certain geographies. Veritone is a global organization, and we, therefore, have team members in various locations throughout the world. These team members often surface information pertaining to certain rules or laws that might be forthcoming and could potentially impact both existing customers and prospects alike. Our legal and compliance products can often help customers satisfy these requirements, and, as such, we’ll then make a formal effort to reach out and assist.

We find that our existing legal and compliance products have applicability to customers regardless of their location, so we don’t necessarily build new products that are solely for customers in a certain geography. Instead, it’s more likely that we tailor solution deployments for customers that have specific data residency or protection requirements due to their physical location.

How does aiWARE work? How do businesses adapt to an operating system to make AI work?

Veritone aiWARE is the first operating system for Artificial Intelligence, making 320+ cognitive engines available through a common software infrastructure. It works by abstracting away the elements of AI that do not directly contribute value to the user so that the user can instead focus on AI’s application to the use case for which they are trying to solve.

A great analogy is Microsoft’s Windows operating system. With Windows, the average user does not need to worry about RAM allocation or CPU optimization; instead, the operating system itself orchestrates those tasks so that the user can focus on their immediate task such as browsing a website or editing a document. The operating system removes the need for the user to worry about things that must occur to browse the website but doesn’t directly provide value to the user.

Similarly, aiWARE orchestrates the tasks removing heavy workload from the users. For example, one component of aiWARE, Conductor, can significantly improve the accuracy of a given cognitive output by running multiple engines, which are specific to an AI capability, such as speech-to-text, – each trained and optimized for a certain purpose – by blending the highest confidence results from each and providing the user with a highly accurate, consolidated output.

In a similar vein, aiWARE future-proofs AI for users, as they are not limited to a defined set of cognitive engines. Instead, they can be added or removed as needed. As new AI capabilities or techniques emerge, they can be easily added to aiWARE so that you don’t need to “rip and replace” and install a new solution.

How can the legal and compliance industry best leverage Veritone’s industry solutions?

Veritone has a number of AI-powered applications that sit on top of aiWARE, purpose-built for legal and compliance customers.

First, Veritone Illuminate is an application that employs a variety of cognitive engines – from Machine Transcription and Translation to Facial Recognition and Text Analytics. It helps legal and compliance customers quickly identify relevant subsets of information from the vast collection of data. For example, customers can use Illuminate to make unstructured data – like audio and video – searchable, by not just keywords and phrases, but also by automatically-identified faces, objects, topics and more. A user could enter a search query such as, “Show me all videos where John Smith appears and telephone recordings where he spoke,” but do so only when he said certain words or phrases. With Illuminate, this can be achieved from a single search query, which is extremely powerful. It also results in immense cost and time savings for eDiscovery teams seeking to cull their dataset down to a relevant subset, as well as for financial services compliance teams that want to identify if material non-public information was shared outside of a control group.

Second, Veritone’s Redact application makes it easy to obscure private or otherwise sensitive information from audio and video evidence by leveraging best-in-class cognitive engines. With Redact, for example, legal teams can easily identify private or sensitive information located within audio and video files within minutes. They can then automatically conceal that information by blurring faces or other sensitive items like license plates and identification cards and insert white noise over confidential segments of audio – a process that previously had been very time-consuming.

Which industries are Veritone’s applications useful to and how?

In addition to the legal and compliance industries, Veritone has purpose-built applications for use cases in the state and local law enforcement, federal and media and entertainment industries. That said, there is cross-vertical applicability to many of Veritone’s applications, and customers can also engage with all of our cognitive capabilities via APIs to AI-enable their existing applications.

What is your take on the raging trend of including ‘AI in everything’ in business?

AI’s use in business is a raging trend for a reason. There are processes that, even in today’s modern age, are entirely or partly manual, and AI can be applied to reduce cost and improve efficiency in these cases. AI lives up to the hype, and folks are beginning to realize this, so there’s much excitement about where it can improve things.

Can you discuss an interesting use case when Veritone’s solution solved a major legal and compliance industry problem?

Veritone was recently contacted by a customer in New York about a forthcoming law that takes effect in January 2020. This new law requires county district attorney (DA) offices in the state to turn over evidence related to a case within 15 days of arraignment. This greatly shortened timeline requires DA offices to collect and sift through massive volumes evidence – including audio recordings and video footage – in an extremely short amount of time, which has led to much consternation about the feasibility of meeting this new requirement by January. Our customers conveyed that they viewed their biggest challenge as not having enough resources to manually review audio and video evidence in time.

Veritone was able to solve this problem by making unstructured audio and video evidence searchable. Instead of manually reviewing this evidence to determine relevancy, DAs can now ingest all of their evidence into Veritone Iluminate, which automatically enables users to search by keywords, faces, objects, and more – making the process of identifying the key evidence much easier and less time-consuming.

We are now solving this problem for a number of DA offices in New York, and we suspect that this will become even more widespread as other states potentially follow New York’s lead with this law.

Which events and webinars would you suggest to our readers as being the best in grasping information on emerging technologies?

I’m a big fan of TechCrunch Disrupt, an event that brings together many of the brightest minds in technology to discuss not only current industry trends but also emerging trends and innovative technology. Disrupt is inspiring and it breeds great ideas. Beyond that, I like to seek out pitch and demo day events that various startup boot camps or accelerators host. It’s a great way to see a showcase of startups that are on the leading (and sometimes bleeding) edge of technology. I would suggest that readers find local boot camps to see when they’re hosting events and then attend them, as they provide great education and are much more interactive than webinars.

What do you think about the ‘Weaponization of AI’?

While I don’t think that we’ll see doomsday scenarios like in “The Terminator” any time soon, I do think that it’s important for the United States to stay ahead of the curve with respect to investing in R&D in this area. Leading nations are looking to AI for a variety of use cases, including unmanned vehicle navigation, autonomous weapons systems operations and finding holes in sensitive systems. Given what’s at stake, it’s imperative that we as a nation place a high priority on investing in this area.

Where do you see AI/Machine Learning and other smart technologies heading beyond 2025?

It’s difficult to distill a single, concise answer to this question. There are so many possibilities, and AI is truly positioned to shape our future. But as AI advances – both in terms of its capabilities, as well as in its adoption – society as a whole will begin to reap the rewards.

For businesses, as an example, AI will become a real competitive differentiator. Whether the use case is increasing revenue, decreasing cost or better understanding the customer, businesses that are reluctant to weave AI into various processes stand to be at a potentially grave disadvantage to their peers that do.

As for individuals – take health care for example. AI has the potential to diagnose diseases like various cancers earlier and with greater accuracy. AI can understand patterns and nuances in data at scale in ways that humans can’t, and we all stand to benefit from this.

What startups are you keenly following?

OpenAI really interests me. I’m not sure that one could categorize it is a start-up at this point, but I think some of the research that they’re doing to advance Artificial General Intelligence is quite fascinating. And, most importantly, they’re trying to do this in an ethical, careful manner.

Tag the one person in the industry whose answers to these questions you would love to read.

I’d love to read Elon Musk’s answers to these questions, given his involvement in OpenAI.