EXCLUSIVE: “The Automated Bank” – Alexey Gabsatarov, Kroo in ‘The Fintech Magazine’

UK challenger Kroo has a mass market brand proposition, but delivering it profitably means you have to run a tight ship, says CTO Alexey Gabsatarov

Market-leading benefits are a sure-fire way to woo customers to your new current account – but generosity comes at a cost.

So, for Kroo Bank, harnessing AI to ensure ruthless efficiency is absolutely crucial.

“The flip side of the coin of being good for our customers and passing on as many benefits as possible to them is we have to be very lean operationally,” says the business’s chief technology officer Alexey Gabsatarov. “Whatever efficiencies we can find, we need to consider them. One way is automating all the manual processes we can, and that means using AI extensively.”

Established in 2016, it took Kroo Bank until June 2022 to gain its UK banking licence. It launched its free current account six months later, swiftly followed by personal lending, and the 100,000 customer milestone was passed by June 23 2023. At the end of last year, it had 152,055 accounts and a combined credit balance of £788.2million.

Ambition and focus

Kroo targets Gen Z and millennials with customer referrals and social media advertising driving recognition – brand building being in co-founder Nazim Valimahomed’s DNA after a lifetime promoting big-name FMCG businesses, such as Coca-Cola and Heineken. There’s a similar vision behind Kroo – to create a highly consumable bank that can nail the rankings.

Valimahomed has already said it’s out to beat Starling’s banking-licence-to profitability roadmap, which took six years. Gabsatarov says the “difficult bit” of creating a current account is now done and additional products will be added as the firm progresses. But while he is confident that profitability can be achieved in the “not-too-distant future”, a £26.8million loss in 2023 means the hunt for operational efficiency must be laser-focussed.

Kroo’s current account pays 4.1 per cent AER on balances by tracking 0.9 of a percentage point below the Bank of England base rate (the margin will rise to –1.1 from late November). That, together with an account app that allows customers to easily split bills with friends, save into pots and enjoy low costs when spending abroad has won the bank praise among pundits.

“We are quite open about the way we use AI and we want to ensure our customers trust us to do it responsibly”

Gabsatarov says: “You may think that a new bank won’t have many manual processes but there’s so much heritage complexity in the way banking operates that automation is a continuous effort for us. It’s important to ensure our AI models run on good quality data – if bias or data quality issues are introduced we see issues downstream.

“From the outset, we’ve adopted ethical AI principles by always having a human in the loop. We’re not replacing humans with AI, the way we see it is we complement specialists with generative AI input that makes them more efficient.”

The bank uses both general and specialist AI systems, says Gabsatarov, as both have strengths and weaknesses.

“Using a large language model can be like interviewing someone who is very smart but who’s not a specialist, so they’ll talk around the topic,” he explains. “It’ll make the answer sound plausible, but ultimately incorrect. Training a specialist large language model counters this problem – you don’t expose it to irrelevant data. We are open about the way we use AI and we want to ensure our customers trust us to do it responsibly.”

Another efficiency win has been reducing customer queries about transactions via a partnership with Snowdrop Solutions. Snowdrop’s software means account holders at partner banks can see merchant names and logos on transaction data, plus the payment location and merchant contact details. It means the bank fields fewer queries from people who don’t recognise a transaction they’ve made and fear they’ve fallen victim to fraud.

Simply adding a logo and location reduces ambiguity and stops misunderstandings from becoming a drain on staff time.

Gabsatarov says: “We launched that a year ago and we’ve seen a 15 per cent increase in customer engagement with that data and a reduction in customer support queries – it’s improved user experience.”

Another user experience Kroo is developing is automated financial advice, or as Gabsatarov explains, “automated nudges and recommendations that would make customers demonstrably better off.”

Neos the world over have adopted similar PFM tools to a greater or lesser extent, but Gabsatarov says Kroo goes further: “People engage with current accounts as individuals but increasingly spend in a social context. So they travel together, live together and spend together. Capturing and serving this element was an important foundational principle for us, and it was a differentiation of ours.

“We’re working on savings products, lending products and so on – adding products that our customers ask for and see value in. Being a small bank we can react to changes and build the business by listening to customers.”

This article was published in The Fintech Magazine Issue 33, Page 11