Open banking startup Abound nabs $601M to supercharge its AI-based consumer lending platform

After a slow start in the UK and Europe, open banking is catching on with fintechs, who are using APIs to access banking data and rails and using them as an alternative to traditional credit networks to build products.

Now, a UK-based consumer lending service called Abound is doubling down on its ambitions in the space with a big fundraise to fuel its own open-banking-based business. The startup has raised a whopping £500 million ($601 million at today’s rates) — money that it will be using to help finance loans, to bring more customers on to its platform, and to invest in its technology, which combines open-banking data and machine learning algorithms to build what Abound believes is a better “credit score” for applicants. To complement its direct-to-consumer offer in the UK, Abound also plans to expand as a B2B service in Europe, which has been building out its own open banking framework, PSD2.

“We see ourselves as going beyond credit scoring,” CEO and co-founder Gerarld Chappell said in an interview, who describes the bank transaction data that Abound uses to build its AI-based risk and lending profiles as akin to “financial X-rays.” These in turn help Abound “understand true affordability” when it comes to loans.

Its rise comes at the same time that we’re are seeing a lot more activity around open banking. Last year, Visa acquired open banking developer Tink, with provides API rails for thousands of banks, for more than $2 billion. Another major rails provider, TrueLayer, last raised at over a $1 billion valuation (granted that was back in 2021…). Meanwhile, Token.io and Vyne are, similar to Abound, examples of startups building more specific applications on open banking standards (respectively person-to-person payments and merchant services).

Abound’s new funding includes both debt and equity: with U.S. bank Citi plus clients of Waterfall Asset Management providing the debt portion; and K3 Ventures, GSR Ventures, and Hambro Perks providing equity.

As is typical with lending startups, the vast majority of the $601 million here is debt, which will be used for lending; the smaller equity part will be used for investing into the business itself. Abound is not disclosing valuation, but for some context, Chappell confirmed that the startup, previously known as Fintern.AI (which is technically still the parent company’s name), had previously raised just under $11 million in equity and around $60 million for loaning. 

More pointedly, the reason for the big sum raised here is that Abound has been seeing surge of interest since launching in 2020.

Its service — based around loans of between £1,000 and £10,000, with repayment options extending up 5 years (although average repayments have been 2.5-3 years), with interest rates the company guarantees are lower than those offered by banks (currently they are 24.8% APR) — has been growing on average 30% month-on-month; it has issued loans to more than 150,000 customers to date; and it says it is on track to loan out £1 billion ($1.2 billion) by 2025.

All of that signals not just something about the state of the economy today, but also the state of fintech. Yes, loans are definitely in demand for average consumers at the moment to supplement their regular monthly income. But it’s also notable to see how new fintech services are being accepted and adopted as a means to getting that liquidity. It’s no longer a novelty to use neobanks and apps to manage money, in other words; it’s just another way, and maybe for some a better way, getting it done.

Chappell said that he his co-founder Michelle He came up with the idea of building Abound years ago when both were working in management consulting — Chappell at McKinsey and He at EY — where they worked with giant financial services clients helping to build credit products and working around the foundations of open banking. The two saw that the API framework presented a clear opening for those who could understand how and where they could be used, he said.

“Consumer credit is very broken,” he said. “Most of it is deeply entrenched in tech from the Seventies and Eighties.”

That entrenchment involves FICO credit scores, and access to that data dominated by companies Equifax and Experian to determine credit-worthiness. Add to this a generally poor consumer experience for loans, and the fact that we’ve seen a lot of exploitation with predatory loan practices, and you can see the gap in the market for better products that address customer needs in a better way.

Ironically, the status quo for loan products actually might be fine for many consumers, in particular, those who have credit histories, he added, and can be clearly categorized as “prime” or “subprime” cases. However, it’s practically unusable for those who are new in the market, so-called “near prime” consumers. There are some 15 million of these in the U.K. alone, Chappell estimates.

Most lenders will reject loan or credit applications from these consumers, he said: “They are just too uncertain.”

So the solution was obvious: build a system that taps open banking to get basic, real-time details about how an individual manages incomings and outgoings in a regular bank account; extrapolate insights from that data using AI; and create a new kind of credit score. This is what Abound took some three years to build before launching in 2020, and it is what now forms the basis of its business.

It might seem obvious that a bank itself could, should, and would come up with something similar to provide its own banking-data-based loan products — at least for its own customers, if not loans for those who bank with others. But Chappell said it’s not as simple as it looks.

“This is very non-trivial. It would take banks five years or more to change their processes,” he said. Their processes typically are those very services entrenched in 70s and 80s methods: FICO scores, combined economics data from the UK’s statistics office, ONS, to determine loan eligibility, as well as how likely it is that a customer might default on a loan, or pay it back as agreed.

Meanwhile, Abound says its proof has been in the pudding, so to speak: Chappell said that in the last two years, the startup has had 70% less default rates than the industry average in the U.K.

That’s to say, it’s not perfect, but appears to work better than what it’s looking to replace. “For every 10 defaults at a competitor, we have three,” he said.

But since this is technology we are talking about, the race doesn’t end when big banks or other startups build what Abound has built: Abound believes it has first-mover advantage, and so, in the time that it will take competitors to build something similar, Chappell believes that Abound will have developed even better AI algorithms to manage its own rates better.

That head start is also what is motivating investors to back the company.

“The lending industry is dominated by old practices, like traditional credit scoring, which ignore technological developments of the last decade,” noted Kuok Meng Xiong, CEO of K3 Ventures, in a statement. “Abound is delivering a unique product and a differentiated approach which is already proving itself to work for thousands of customers. We are excited to see Abound’s offer grow in the years ahead.”

“Waterfall is pleased to be part of Abound’s business expansion as it seeks to utilize open banking in a more informed way to help the consumer,” added Krishin Uttamchandani, director at Waterfall Asset Management, in a statement. “Abound is led by a strong management team that we are excited to work with, supported by, what we believe is, a robust tech stack, underwriting methodology and view on risk. What Abound has achieved in its first two years of lending has been very impressive and should lay the groundwork for a strong platform to better serve the customers who should be able to access cheaper credit with open banking. We are excited to be a partner in the Abound journey.”

Open banking startup Abound nabs $601M to supercharge its AI-based consumer lending platform by Ingrid Lunden originally published on TechCrunch