Gradient Labs

Gradient Labs is a London-based company that builds AI agents for customer operations in financial services. Rather than a single general-purpose assistant, it offers a suite of domain-specific agents — including a Lending Agent covering the borrower lifecycle from missed payment to outbound collections calls and agreed repayment plans, a Disputes Agent handling intake, evidence gathering, investigation and chargeback filing, and a KYB Agent running identity and document checks — plus a platform that coordinates the operations between them. Other advertised use cases include customer service (for example freezing a lost card or tracing a missing payment), onboarding and activation outreach, and insurance claims intake.

The agents operate across the channels a bank's customers use, including voice, and are designed to execute long-running, regulated workflows end to end: gathering evidence, deciding outcomes, escalating to a human for sign-off where required, and writing back to core systems. Agents share context and memory with one another and can interoperate with a customer's existing AI agents. Domain-specific guardrails, compliance checks and test scenarios are embedded per agent, with reference to regimes such as FCA Consumer Duty and the EU AI Act.

The company positions itself as a vertical AI provider, arguing that horizontal, general-purpose AI cannot address the operational complexity of banking. It states it has spent three years pursuing full autonomy rather than a co-pilot model, and reports customer satisfaction scores above those of human teams in its deployments.

Founding story

Founded in 2023 by Dimitri Masin, Neal Lathia and Danai Antoniou, who met at Monzo where they led AI and data efforts. Masin was Monzo's 20th employee and led a data team of over 100; Lathia built Monzo's machine learning infrastructure; Antoniou led creation of a fraud detection system. Their experience of the operational burden of running a bank at scale, combined with the emergence of frontier AI models, led them to bet on vertical AI for financial services.

Business model

Enterprise B2B software sold to financial institutions, deploying AI agents into their customer operations, compliance and back-office workflows. The company states that once a use case is scoped it guarantees the deployment, offering customers their money back if it does not deliver what was promised.

Not specified in the sources beyond enterprise deployments with scoped use cases and a money-back deployment guarantee.

Traction

Named customers include Wise, Monzo, Zego, Pockit, Plum, Yonder, and US firms Current, Stash and Rho. The company reports 900% revenue growth over the prior year (described elsewhere as more than 10x), agents supporting over 32 million end users, hundreds of thousands of voice interactions per month in live environments, CSAT of up to 98%, and resolution rates of roughly 40-60% at launch.

Latest developments

In June 2026 the company announced its Series A had been increased to $26 million, led by Octopus Ventures and CommerzVentures with follow-on from Redpoint Ventures and Exceptional Capital, to expand in the US, scale agent infrastructure and cover more financial services workflows. It also shifted from a single agent to a suite of specialist agents (Lending, Disputes, KYB) and, in August 2026, launched Collaborate, a capability for operators, engineers and agents to iterate on agents with version control, evaluations and continual learning.