Fintech Interview: Part 11 AimBrain
The latest installment of my Fintech Interview series features biometric identity management software supplier AimBrain.
The story began four years ago when an interim at chipmaker Arm got inspired by a project which had been scrapped after the research and development phase.
AimBrain is currently targeting finance firms as customers and is based in the Level39 start up community in London’s Canary Wharf. It is a company full of Phds pushing the boundaries of deep learning.
AimBrain has an identity platform that uses multiple biometric technologies to provide user authentication, which is integrated into a workflow at a customer. AimBrain is for example working with a credit card company in the US to help stop fraud related to new credit card applications to stop fraudsters using other people’s details to get a credit card.
“Everything is moving to online and mobile channels and fraudsters are targeting this. At AimBrain we are trying to reduce fraud and improve user experience at the same,” Andrius Sutas, CEO and co-founder, told me.
The company’s biometric identity management as a service platform aims to do just that. It has five modules: facial and voice identity, aggregated face plus audio, behavioural authentication – which tracks multiple characteristics such pressure and range of movement on devices and peripherals, and anomaly detection, which looks for patterns of fraud before the user profile is known.
Behavioural biometrics can be used in banking apps to add security. For example after a person has logged in using facial recognition it can do further behavioral check before approving them to make payments etc. The system will either let them continue or ask them to re-authenticate.
Sutas said AimBrain is as much an identity company as it is a biometrics supplier.
He previously did an internship at chip maker Arm after graduating with an economics degree. He said it was while doing this temporary role at Arm, in the research and development department, he had the idea.
“Arm was working on a smartwatch and the biggest challenge they had was not creating the smartwatch itself but how to actually identify the user,” he said.
“The problem is a watch doesn’t have a keyboard or a big screen so passwords would be difficult. Arm started looking at biometrics but it looked at the market and could not find anything good enough to buy off the shelf.”
“So it started hiring Phd students to build the capability in-house. But in the end the whole project got scrapped.”
At that point Sutas and best friend Alesis Novik, AimBrain co-founder, started looking at what is available in the biometric identity market and what was needed.
“Biometrics has been around for 30 years but it has only recently taken off because of companies like Apple, Samsung and Huawei,” said Sutas. Then the payments services directive requiring stronger authentication, with biometrics strongly recommended by many, came in.
Sutas said there was an opportunity to take biometric authentication to a new level. “Until recently biometrics was not that good. It was not accurate and it was slow because machine learning science has not advanced to the point that it could build biometrics.”
When Novik dropped out of PHD in machine learning he and Sutas began working on the concept of AimBrain which was to apply biometric base identification technology to the challenge of managing an individual’s identity. “We are not just a biometrics company we are about identity,” said Sutas.
“Identity is not just about what devices you have, the way you talk or look it is everything combined.”
Sutas said biometric identity needs to offer multiple formats. “When people ask me what is the best biometric authentication to use there is no clear answer because not a single module on its own is good enough.
If you have face recognition it does not work in the dark. Voice recognition does not work when it is noisy and behavioural is invisible to end user and is not as accurate.
“We wanted to create a platform with multiple modules,” he said.
But the first port of call was money. This is where Entrepreneur First came in. This is an accelerator with a bit of a difference, said Suta, as it invests in individuals that have good ideas.
“They benchmark you in your capability and invest so you can build and test something and if it is a success will invest more in the company to take it forward and take a stake in the company
AimBrain is in the Level39 startup community in Canary Wharf, which puts it in and around the community of businesses it is targeting the most with its products, the financial services sector.
This proximity to the potential customer base is vital. Sutas said customer engagement began well before the company had a product to sell. “The theory that when you build something customers will come does not work in reality. We were engaging with before we build anything. It is very important to have someone guiding you when you start.”
He said engagement with the market is vital when developing something new. “It is very easy to go out and build technology without talking to potential customers but then when you do go to them you find out it is not what they expect.”
Sutas said it is also vital that you speak to the right people within customers, which he believes should be the business units that will use the products.
Many banks are looking for fintechs and hire people to bring in startups to their incubation programmes. But the people that bring them in, the innovation teams, are often doing so to showcase technologies not necessarily be ready to use them
It is better for fintechs to talk to business units that are looking for a solution to a problem.
“You have to do anything you can to get to meet these people. This could mean waiting outside the office for them or inviting them for a beer,” said Sutas.
This customer engagement is helped by AimBrain’s membership of Level39. “It is an ecosystem. Not only does it introduce you to various potential customers but puts you in contact with other companies at Level39 that can use our technology within theirs”.
AimBrain currently processing over 100 million authentication requests a week. It has 20 staff, more than half are techies with PhDs.
It is a small company but its team are at the leading edge of technology development, including deep learning, with patents in the machine learning fields. “We don’t just use machine learning just as a product we have algorithms training other algorithms on biometrics,” said Sutas.
In the future Sutas wants AimBrain to be everywhere but invisible to users like the products of the big chip makers, such as Arm.
“We are not a standalone product but have to be integrated as part of a workflow. What we want to create a brand similar to Arm. It will be built in everywhere but no consumers will know about it.”
Read the previous fintech interviews
Part 10 Meniga, Part 9 TrueLayer, Part 8 InvestCloud, Part 7 ClauseMatch, Part 6 Rebuilding Society, Part 5 Honcho, Part 4 Akoni, Part 3 Wrisk, Part 2 CreditLadder, Part 1 Taina Technology