IT meltdowns are becoming dangerously common among the big banks. According to Which, at least one British bank suffered an IT failure every day during the last nine months of 2018.
TSB’s outage last year, which left 1.9m customers locked out of their accounts and cost the bank £330m, demonstrates how catastrophic these IT issues can be to both businesses and consumers.
The problem for many banks revolves around their use of outdated legacy technology, which often buckles under the weight of modern-day demands.
Patching up the problem only delays the inevitable collapse, and many banks are now accepting the need to take bold action.
“Legacy software isn’t getting any better, and the gap between the old and the new is increasing, while customer expectations are also changing,” says Paul Taylor, founder of Thought Machine, a fintech firm which is building retail banking platforms in the cloud.
Launched in 2014, the startup has already partnered with Lloyds in an £11m investment deal, as well as Atom Bank. I’m also told that more announcements are expected this year.
“It’s a fresh start for the banks – freedom for the legacy,” Taylor says.
Head in the cloud
From where I’m sitting, Taylor seems way ahead of his time. He completed a PhD in machine learning years before everyone was talking about artificial intelligence, and he lectured on speech technology at Cambridge University as far back as 2000.
A self-professed geek (he had a BBC Micro computer as a kid), it’s funny to hear him say that writing software was considered a specialist hobby back in the eighties.
“I used to write video games and play in a band, and my parents said I might make money out of music, but that I would never make money out of playing with my computer,” he laughs.
Now, of course, it’s hard to imagine our world without software – it’s had a huge knock-on effect on every aspect of our lives, including the way we manage our money.
Many banking customers will probably be grateful that Taylor stuck with the software.
Banking on change
Even for a man who is responsible for the text-to-speech technology used on Android phones, Taylor admits that revolutionising retail banks is not easy. “Everyone in the industry said that the idea was nuts – they said it was too complicated,” he remembers.
When you’re dealing with millions of banking customers, there are no small sums of money at stake. But the mass and wealth of customers isn’t the only issue. The existing IT infrastructure – which is often clunky and complex – poses its own problems.
“We visited a bank recently with a legacy system, and they had no idea what everything did, because the people who wrote the software have all left the company. If they wanted to turn it off, they wouldn’t know how.”
A brave new world
While Thought Machine is part of the fintech movement, the company is not focused on providing additional services to banks (as many fintechs are), but completely replatforming the financial institutions.
I ask whether Taylor thinks that some fintechs might have underestimated how complex these banking systems are. While that might have been true, he says that the dynamic is changing. “In recent years, the relationship between fintechs and banks has matured – everyone realises that scalability, resilience, and security have to be taken very seriously. There’s definitely a meeting of minds.”
But in many ways, those operating in the big banks and fintechs are from different worlds, and Taylor admits that he doesn’t have much direct knowledge of the traditional banking systems. “One of the senior guys at Lloyds asked me to compare and contrast the new and the old, but I only really understand the new. They tell us all their problems, and we explain our solutions.”
Thought Machine isn’t just helping the conventional, established banks – evident with its Atom Bank partnership. Taylor tells me that challengers have been disappointed by the lack of options around banking software.
“Building software themselves was possible, but it takes a long time and is very expensive. And there is also no point starting a new bank with old technology.”
The day after tomorrow
So how does Taylor make sure that his company isn’t getting left behind in this fast-paced tech world?
“Until relatively recently, the industry has made a bad habit of only updating their software once a year, and then they fall down exhausted afterwards. But new things are coming along all the time, so we release a new version of our software every two weeks to our banks. And once you get into the rhythm of that, it changes your mentality.
“Small changes are easier to absorb than big changes, and once you get into the rhythm of making those small changes, it’s less disruptive.”
It’s clear that many financial institutions have been complacent. “You have to make sure that you aren’t building tomorrow’s legacy,” says Taylor. “You have to build the software with knowledge that it’s going to change.”
Documenting and testing every piece of code is a practice which was drilled into Taylor and the senior members of his team while working at Google (Taylor co-founded speech tech company Phonetic Arts, which was bought by Google in 2010).
“Google has pushed the forefront of what’s possible in technology – and not necessarily in terms of what the consumer sees, but how the software is deployed,” he says.
“When you’ve been through the Google machine, you can’t go back. So when we started this company, we wanted to do our own thing, but we have adopted those practices.”
Pre-internet revolution, Taylor used to joke about whether his software would actually be used in a day-to-day sense. But now, software companies like his are building the future. Poignantly, he says: “We owe it to ourselves and to society to get it right.”