He may only be 23 years old but Russia-born Dmitry Aksenov has certainly packed a lot into those years. As a precocious, technically gifted seven year old he built his first autonomous robot, triggering a lifelong passion for robotics and artificial intelligence – he’s completely self-taught in the fields, winning national and international robotics competitions while still in his youth, even representing his country at events at one stage.
And now, as founder and chief executive of DigitalGenius, Aksenov is using his technical wizardry to drag AI out of the lab and into the real world, focusing first on helping big companies provide a better, more personal service to billions of their customers.
The company’s flagship AI solution, FinGenius, combines the latest advances in artificial intelligence – the science and engineering of making intelligent, self-learning computers and software – with cutting edge natural language processing to create technology that can answer questions from customers in any language without the need for human intervention, dramatically cutting the need for conventional help desks and call centres.
FinGenius’s core innovation is in “deep learning” – a branch of AI focused on enabling computers to learn and “think” by themselves. It involves creating “neural networks” – software that tries to mimic the way biological nervous systems, such as the brain, processes information.
Aksenov explains: “Specifically we are focusing on conversational AI – our core, proprietary technology is all about understanding natural language so that we can tackle conversations within big businesses as well as between them and their customers. It can receive a question in any natural language through any communication channel, whether it’s voice, text or email. It understands what that question means and provides an answer to it.
“It’s a self-learning system that is specially designed to automate the conversational experience for big brands and their customers. It can really help with call centre operations of any bank or Fortune 500 company – financial services is a big big target for us, but we work with other industries as well. The major challenge for these companies is that their call centres get a lot of customer calls and they’re expensive to deal with in volume. AI can help with that”
Learning on the job
In practice the FinGenius technology – which operates on-premise or via cloud services – can read logs of previous conversations between customers and company and learn from them. It can also learn on the fly, monitoring and learning from conversational exchanges in real time. The system can answer thousands of questions simultaneously with tailored responses for each individual customer. After being integrated into a company’s call centre operations the end result, typically, is the automation of responses for a huge number of them. Aksenov says that up to 80% of phone calls in the financial services sector fall into just five categories. These include people phoning to notify that their card is lost or stolen; that they’ve forgotten a password, or can’t make a certain transaction: “These are fairly standard phone calls – it’s very easy to automate all of them.”
The remaining 20-30% of calls – the trickier, more complex ones – are passed on to a human by the FinGenius system. “We are not aiming to automate the entire customer service experience. We’re just automating the most common questions. We believe that human and AI should work together. AI is not a panacea. It only empowers live [call centre] agents to focus on things that can add way more value, give them more time to, for example, focus on outbound sales instead of answering repetitive questions in large volume. It is about providing a better customer experience while also generating more sales and significantly reducing costs.”
Not surprisingly given the potential benefits from its application, FinGenius has attracted considerable corporate interest. DigitalGenius, which employs 29 and has offices at London’s Level39 and in New York, is working with dozens of mainly UK and US-based companies to explore the potential application of FinGenius in their operations. Some have been fully onboarded as clients. They include BMW, Panasonic and Unilever.
The company is also working with many financials, including eight banks the company struck deals with through its participation in Accenture’s 2014 Fintech Innovation Lab start-up mentoring programme at Level39. “Most of our client relationships – almost all of them are Fortune 500 companies – are subject to non-disclosure agreement so we are unable to names. But we are working with a lot of Tier 1 banks, mainly retail. We are at different stages with those particular relationships – some are still in pilot stages, some are live.”
Very recently DigitalGenius decided on a major change in strategy: the two and half year old company is now transitioning from a project-based enterprise, which Aksenov says limited the number of companies it could work with at any one time, to one that is product-based. It’s a move that promises to help accelerate growth: “We are getting ready to launch an off-the-shelf product that can be purchased without any professional services. As a result our sales cycle will shrink significantly. Instead of spending two months on a single deployment, which was causing bottleneck problems, it will take just a few days to deploy our technology. With the new product-based approach, we can pretty much work with as many clients as we want. There is still a limit but it is a lot higher than it was before. ”
The company is gearing up for a full blown sales campaign for the new product in the first quarter of 2016 and Aksenov is hoping it will lead to many more clients being secured.”
Compliance: the big DigitalGenius target
While the current priority for DigitalGenius is helping corporates – most especially financials – communicate more efficiently and effectively with their customers, Aksenov is clear his team has its sights on other applications for the FinGenius technology going forward: “As we grow we will be expanding into other use cases. Some of these are focused on generating internal efficiency gains for organisations: how can we simplify the day- to-day experience for employees? Can we make it easier for employees to find information and answer questions?
“The next one – the really big one for us – is compliance [in the financial sector]. How can we use our conversational AI technology to make sure that every outbound communication is fully compliant, obeys all the rules? How can we use it to prevent, say, banks from miss-selling PPIs?
“Compliance is too big for us to work on right now because sales cycles are really long in financial services industry and we don’t want to spread ourselves too thinly at the moment anyway. Also the solution will need to include some professional services component, whether from us or from a company like Accenture. Once we get to certain points, milestones, possibly around second quarter of next year, we will start looking at compliance. It will need to happen.”
Aksenov’s ambitions are ones that Derek White, chief design officer at Barclays, would no doubt applaud. In a recent interview with Fusionwire.net White says there is big demand for new technology that automates functions and removes the need for paper. More broadly, says White, there is a need to improve the lot of people who work within institutions by making “their day to day experience richer, much more rewarding”. He adds that such areas are ripe for disruption through, for instance, artificial intelligence, natural language processing and hands free computing as well as leveraging of the cloud.
The growing interest in FinGenius’s solutions from corporates has been matched by that of investors keen to tap into DigitalGenius’s potential while it’s still at an early stage in its development. In June the company carried out its first major fund raising, securing $3m in a round led by Metamorphic Ventures, with participation from Lerer Hippeau Ventures, Lowercase Capital, RRE Ventures, Lumia Capital, and a number of angel investors. Lowercase is particularly noteworthy, run as it is by Chris Sacca. He made his billions through early bets on Twitter and Uber and a number of other start-ups. He was also one of the first investors in Instagram, acquired by Facebook in 2012 for $740m.
Aksenov says of the DigitalGenius investor base: “It’s a very, very strong group I think. As well as Lowercase we have big funds like RRE and Lumia, who typically invest in much later-stage companies but still decided to go with us as they saw an opportunity for the long-term and they feel comfortable with us, our plans.” Further fundraisings are planned, he adds.
AI on course to deliver at last?
Born in Vladikavkaz, located at the foothills of the Caucasus Mountains, Aksenov came to the UK in 2008 to study at a further education college before taking up a business administration degree at Bath University: “It was quite exciting for me because I had been involved with technology all my life. I am a technology geek, I read a lot of books and did a lot experiments when I was a kid. But now I wanted to learn something different so I decided to learn business to broaden my knowledge. It was a really fun experience.”
Having been so immersed in AI throughout his life Aksenov is eminently placed to comment on the current state and prospects for a field that has promised much for many decades but never quite delivered. Things look very different this time round: “Well, the technology required is definitely getting better and better and better – the computer industry now understands the type of hardware we need in AI. The processing power, cloud that has become available in the last few years has helped a lot also,” he says.
Another major driver behind the surge in AI activity globally in recent years is the improvement in the quality of research being carried out and published in the field. “It is the type of research that we need to get closer to emulating the way the human brain works. It really helps that companies like Facebook, Google and Microsoft are opening huge, enormous AI labs. They’re making their research available publicly which helps smaller companies like ours, we benefit from it. It’s great that some of the smartest minds in this space are cooperating with each other to build something great. It’s not knowledge locked within any one physical organisation, it’s being shared.”
Yet Aksenov is keen to manage expectations: “AI is not going to revolutionise everything in the next year. AI will not replace everything in the next five years. It will not be able to replace a call centre. It will not be able to replace fundamental processes. What it can do is assist humans. So, it can work together with a person and give them super-powers, in a way. So, instead of spending hours and hours crunching some data, you can give this repetitive stuff to AI. Instead of answering the same questions over and over again, you can give the stuff to AI. Instead of going through a lot of logs to spot any compliance challenges you can ask AI to do it. There are certain things where AI is better than a human. There are certain things where a human is better.
“In the next few years, we will see a lot of collaboration and cooperation between humans and AI. A lot of repetitive, boring stuff will all be outsourced to an AI, to a machine. In five years or so, within that sort of timeline, we will get closer to more generic AI where a machine can start to understand how the world works around us and we’ll be able to solve a lot more challenging questions.”