Companies in areas such as financial services and insurance live and die by their data – and specifically how well they can use it to understand what people and businesses will do next, a process increasingly dominated by AI. Now a startup called Finbourne, founded out of the financial center of London, has built a platform to help financial companies organize and use more of their data in AI and other models. It is announcing £55 million ($70 million) in funding, which it will use to expand its reach beyond the Square Mile.
Highland Europe and AXA Venture Partners (also known as AVP, and backed by the insurance giant of the same name) are co-leading the Series B, which values the company at just over £280 million ($356 million) post-money.
Thomas McHugh, the CEO who co-founded Finbourne, told TechCrunch that he came up with the idea for the startup after working as a senior quant in the city for many years, most of which was at the Royal Bank of Scotland. One of those years was 2008, the year when RBS, then the world’s largest bank, dramatically came to the brink of collapse after being overexposed to the subprime lending contagion.
The major shift took place internally in the form of a massive reorganization.
Previously, the entire bank was organized into a series of business silos, which affected not only the way people worked, but also how the data within them worked. All of this cost a fortune to operate, costs that urgently needed to be cut. “We had to remove hundreds of millions of costs from the company in a very short period of time,” he recalls.
They decided to take a page from the emerging but fast-growing world of cloud services. Founded in 2006, AWS had only been around for two years at the time, but the data teams could see that it presented a compelling and comparative model for how a bank could store and use data. Therefore, a consolidated and federated approach to the problem was also needed.
“We managed to build a ton of technology that worked across every asset class. Until then, people said this wasn’t really possible. But we had an incredible reason to change and because of that we knew we could build better technology, much more scalable technology,” said McHugh. Equity systems, fixed income and credit, he said, which were previously all managed as separate systems, were now on one platform.
The 2008 British financial crisis was a rollercoaster that, if you hadn’t been thrown off completely, you certainly wouldn’t have believed you could survive and overcome any challenge. So that ultimately led McHugh to take on the riskiest of all things in business: a startup.
Finbourne may have its roots in the way McHugh and others in his team took on the challenge of building more efficient data services at their bank, but it also developed the idea that reflects and shapes the way financial services companies today day buying IT. Just as companies with extensive sales operations might use Salesforce or a competing platform instead of building their own software, Finbourne expects that financial companies will increasingly do the same: partner with third-party companies for tools to run their operations instead of their build your own software.
This inevitably also ties in with the way in which banks and others in the financial services industry are increasingly working with AI.
Today, the company’s products include the LUSID Operational datastore; investment and accounting records (used in asset management analysis); a portfolio management platform that tracks positions, cash, P&L and exposure; and a data virtualization tool. McHugh said Finbourne also helps manage how companies handle their data for training models, an area where it is likely to become more involved.
It seems like the main takeaway here is that there is no clear leader, and banks don’t want to share data with other banks, so they train ways to prevent that – a process that also helps customers control results more tightly and prevent “hallucinations” creep into the picture. Open source plays an important role in how it provides more flexible options to end users.
“What we’ve seen is that customers don’t want the models we write or use to be trained on someone else’s data,” he said. “We see that very strongly. We do it because these models are less able to hallucinate because they are not allowed to use someone else’s photo.”
Finbourne currently has a slew of competitors. Rivals in the asset management space include Blackrock’s Aladdin, SimCorp, State Street Alpha and Goldensource; Other competitors in the alternative asset manager space include Broadridge, Enfusion, SS&C Eze and Maia. BNY Mellon Eagle, Rimes, Clearwater Analytics and IHS Markit all offer tools for asset owners; and asset services include FIS, Temenos, Denodo, SS&C Advent and NeoXam.
The fact that there are so many could be a compelling reason for someone to take a more simplified approach by working with just one person – a route that companies like Fidelity International, the London Stock Exchange Group, Baillie Gifford, Northern Trust and the Follow Pension Insurance Corporation. PIC) are taking.