The bet on AI-powered customer service just got bigger for a trio of former Monzo executives who think they've cracked a problem that has long vexed financial institutions: automating the messy, regulation-heavy interactions that define modern banking.
Gradient Labs announced June 1, 2026 that it's closed a $13 million extension to its Series A, bringing the round to $26 million total. Octopus Ventures and CommerzVentures led the extension, joining earlier backers Redpoint Ventures and Exceptional Capital. The London-based startup, which opened a New York office last fall, describes the resulting investor lineup as deliberately transatlantic—a cap table that mirrors the geography where it's racing to build what co-founder Dimitri Masin calls "the foundation for autonomous banking."
It's an ambitious phrase. But the numbers suggest something is working.
From Monzo's Data Trenches to AI Agents at Scale
Masin and his co-founders—Neal Lathia and Danai Antoniou, both veterans of Monzo's data and machine learning teams—launched Gradient Labs in 2023 with a focused thesis: banks and fintechs could deploy AI agents to handle customer operations, but only if those agents could navigate the thicket of regulations that make financial services so operationally complex.
Since then, the startup claims revenue has grown 900% year-over-year, fueled by deployments at Wise, Monzo, Current, Stash, Rho, Zego, and Pockit. The platform claims to now touch more than 32 million end users, handling workflows from loan applications and collections to disputes and KYC verification.
Voice has emerged as a breakout channel, perhaps more than the founders initially expected. Gradient Labs reports processing hundreds of thousands of customer calls each month, with its Borrower Lifecycle Agent alone fielding upward of 100,000 calls monthly across the customer base. An April customer story published by OpenAI highlighted resolution rates exceeding 50% on day one for complex cases, with satisfaction scores reaching 98% in some deployments.
Those metrics—if they hold—would represent a meaningful shift in an industry where call center experiences have historically ranged from merely frustrating to outright Kafkaesque.
Building Guardrails for a Regulated World
What distinguishes Gradient Labs' pitch, at least on paper, is its focus on compliance-first architecture. The platform ships with pre-configured guardrails designed to satisfy regulations including the UK's FCA Consumer Duty, the EU AI Act, and a litany of US statutes: FDCPA, TCPA, Regulation F, UDAAP.
It's a practical necessity. Financial services remains one of the most heavily regulated sectors globally, and any AI misstep—an agent that overpromises, mishandles data, or violates consumer protection rules—can trigger fines, lawsuits, or worse.
The technology itself runs on a multi-model stack, pulling from OpenAI, Anthropic, and Google depending on the task. An April product update touted voice latency around 500 milliseconds, low enough to approach the rhythm of natural conversation. The system handles outbound interactions across voice, SMS, and email, orchestrating multiple agents when workflows require it.
Still, the real test lies ahead. LHV Bank, an Estonian lender, announced a proof of concept with Gradient Labs in March, signaling interest from traditional regulated banks—a tougher sell than the neobanks and fintechs that dominated the startup's early client roster.
Timing the Market, or Riding the Hype?
The fundraising environment for AI startups remains frothy, even as venture investors grow more selective elsewhere. Gradient Labs announced its initial $13 million Series A on July 8, 2025, led by Redpoint Ventures, just eleven months after raising a £2.8 million seed from LocalGlobe. The latest extension brings total capital raised to roughly $30 million.
That's modest by the standards of large language model developers burning through hundreds of millions, but substantial for an application-layer company still proving product-market fit across borders. The company now employs more than 40 people split between London and New York, offices chosen to serve what it sees as the two most important fintech markets.
The US expansion, announced in October 2025, has already yielded clients including Current, Stash, and Rho. Whether that momentum translates into broader adoption among American incumbents—the JPMorgans and Bank of Americas wary of ceding customer touchpoints to third-party AI—remains an open question.
For now, Gradient Labs is making a straightforward wager: that banks will choose autonomy over headcount, and that the firms offering compliant, reliable AI agents will win the infrastructure layer beneath autonomous finance. The next twelve months should clarify whether the fintech world agrees.
