A startup called Riff wants to automate where your money sits overnight. Jefferies thinks it will work, eventually, and the consequences for UK banks are severe.
The pitch is simple enough. Riff, which plans to launch the UK's first AI savings agent this year, would use Open Banking and variable recurring payments to move customers' idle current account balances into higher-yielding savings products, automatically, without any action from the customer. The company's founder outlined the model on a call with Jefferies analysts this week.
What the analysts took from that conversation is the subject of a research note that should make uncomfortable reading for anyone holding UK bank shares.
The mechanics of the model
Riff's rollout has three stages. In the first, the agent manages current account balances, sweeping excess cash into a savings account operated through a Banking as a Service provider. Because Riff does not hold a banking licence, the rates it can offer at this stage may be constrained.
The second stage distributes funds to partner banks, at which point rate optimisation becomes the core product. The third stage, which the founder believes could arrive within a year as Open Finance matures, extends the agent's reach to non-payment savings accounts. The company plans to sell direct to consumers but expects white-labelling arrangements with digital marketplace partners to be the main commercial channel.
What the numbers look like
Riff's central scenario is stark. In a market where all customers use savings agents, the company believes 0% balances in current accounts could fall from roughly £500bn to below £100bn. The residual balance reflects only the cash needed to cover typical spending over 10 to 40 days.
On that assumption, deposits would eventually be priced like wholesale funding, with no margin premium for the "costs" of running a deposit-funded franchise. Rates would settle around SONIA.
Jefferies ran the numbers. Current deposit margins of around 190 basis points would compress to around 20 basis points. That translates to a reduction in industry-wide revenue of approximately £40bn, a figure the analysts describe as "not dissimilar to the system's profits."
What banks do next
Riff's founder does not think banks will simply absorb the hit. His expectation is that lenders will widen mortgage margins and cut costs aggressively, with cost-to-income ratios falling below 30%. The net result, in his view, would be sector returns on tangible equity back in the low double-digits, similar to 2019, when deposit margins were also close to zero.
He is less concerned about regulatory friction than many observers might expect. Consumer duty obligations, he argues, will make it difficult for banks to obstruct or delay the adoption of savings agents. On mortgage margins, he believes a lower neutral base rate would offset any political pressure from higher borrowing costs for consumers.
The timeline is the uncertainty
Riff's own projections put savings agent assets under optimisation at between £100bn and £200bn by 2028, representing 5% to 10% of household M4, the broadest measure of money supply. The founder acknowledges that adoption among older customers will be slower and that trust is a central variable.
Jefferies is cautious about the pace but not the direction. The analysts expect adoption to come in waves, particularly during rate upcycles, and they note that those who switch are unlikely to reverse course. Younger customers are more receptive from the outset.
What this means for bank valuations today
The Jefferies view is measured. The bank estimates that around 60% of domestic banks' market value will be returned to shareholders through distributions in the next five years. Even if Riff and its peers reach £100bn to £200bn in assets by 2028, that initial wave may be dominated by deposits that are already low-margin, limiting the immediate revenue impact.
The conclusion in the note is that the threat is real, but the medium-term damage will be modest. The longer-term picture is harder to dismiss.