Albert Brown, senior vice-president of engineering at Veritone, talks AI, automation and ‘in-the-loop’ employees.
Albert Brown is senior vice-president of engineering at California-based AI firm Veritone. His experience in technology spans two decades, having led engineering teams in such companies as IBM and FileNet.
Here, he discusses Veritone’s vision for digital transformation, AI and automation.
‘Data scientists, engineers and business units need to work closely together to develop, deploy and maintain a successful AI strategy’
– ALBERT BROWN
How do you drive tech strategy with AI?
When deciding a tech strategy, it is important for businesses to take a step back and identify the consistent pain points across the business. Next, evaluate what kinds of data your organisation can collect, then refine that data to a quality that is trustworthy.
Once that is accomplished, bring a few hypotheses to the table and discuss how AI can be used to solve these common problems. It is crucial to first think about the pain points, then the data and after that, testing and automating before implementing the tech strategy.
At Veritone, we realised that it was a waste of time for employees to pull out a badge to unlock the office doors each day when they arrived for work. We evaluated what data we already had on hand – employees’ faces – as well as which cameras we could install.
We set standards to unlock the door when it recognised the face of an employee, and we now have an automated system for unlocking office doors for employees only.
A similar process can be applied to many industries. Consider running a manufacturing plant, for example. It is important for manufacturers to ensure employees follow certain protocols, such as loading shipments or using equipment correctly.
Businesses can set rules for proper processes, then run sensor and video data through AI to recognise if employees are following these processes and ensure workers follow set protocols.
How do you help drive IT and tech initiatives at Veritone?
At Veritone, we talk about loops, where humans are either ‘in the loop’ (manually performing mundane, routine tasks), ‘on the loop’ (AI performs the mundane, routine work so humans can move on to the cognitive tasks and provide final review and approval), or ‘out of the loop’ (where humans are removed completely and AI performs all tasks).
We try to identify the tasks where people are in the loop, with the goal of providing them with the proper tools and technology so that they can be on the loop and leave the process, only having to supervise. Eventually, we want to reach a state of advanced technology where humans are out of the loop completely and the AI and automation work independently.
When we see common problems repeated across businesses and industries, where humans waste time on mundane tasks in the loop, we create solutions for those specific business cases and package them up for a number of businesses to use.
How big is your team? Do you outsource where possible?
The size of my team depends on the state of the project. First, I look at the hypothesis and evaluate the level of data and expertise we will need to solve it. If the next step of the process is to prototype, then we will need engineers, programmers and data scientists on the applied end.
At Veritone, we don’t want to recreate the wheel, so we make a point of identifying where a similar solution has worked in the past and if it already exists. If someone has already conducted the research, we don’t need to start with data researchers. Instead, computer science professionals are pulled in to help scale the project. If it is something that has never been done before, we will need to start with data researchers and then build the entire solution ourselves.
Problems and team make-ups will vary based on what the business needs and, as a result, it’s crucial that tech strategy is agile. Businesses need to identify an audience, decide if they are serving innovation groups or IT teams and then build the team out from there.
Tight cross-functional collaboration is going to be the key to success. Data scientists, engineers and business units need to work closely together in order to develop, deploy and maintain a successful AI strategy which may require many organisations to rethink their current team structures.
Video and audio #evidence is piling up in the age of body-worn cameras. Join us for a webinar with @Microsoft and Pemberton PD to discover how #AI can streamline the redaction process: https://t.co/j8QhOUp3Qy @MircosoftPSNS @Microsoft_Gov #MicrosoftGov #redact #VeriRedact pic.twitter.com/3n7NilXi0M
— Veritone (@veritoneinc) March 11, 2020
What are your thoughts on digital transformation and how are you advising customers to address it?
Veritone accelerates digital transformation by supplying the foundation to organisations to digitally transform at scale.
For example, in 2016, during the election campaigns, Veritone aggregated intelligence about what was being said at scale across the US for competitive insights. Broadcast TV folks used this data to track the competition and compile summaries of campaign updates and common trends to better match up against competitors during the ever-evolving campaign season.
Today, we are seeing accelerated adoption of similar technology in the use of sports stadiums, as fans continue to want to engage with the team and the game through the use of emerging technology.
What big tech trends do you believe are changing the world and your industry specifically?
In the next year, technology will continue to rapidly accelerate. We will see a dramatic increase in the adoption of AI solutions as a part of the business. One place in particular where I predict we will really see growth is in smart speakers throughout the home and office.
Currently, smart speakers are automating music and a few things around the home, like lights. Voice-enabled AI will start to expand its usage and allow smart speakers and other pieces of similar technology to be leveraged across the home but, more importantly, at the office for several business use cases. Smart speakers will play an important role in business processes, and we will start to see these speakers as assistants in our everyday life.
Furthermore, AI will become even more intertwined in our lives, handling ever more complex scenarios. Flo, an LA-based company, has an IoT device on water main lines to determine if there is a leak. When it finds one, it automatically shuts the line’s water off. This is an example of smart AI solving critical societal problems, and although the technology doesn’t think for itself yet, it learns from behaviours and finds a solution without involving humans.
Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.