Gradient Labs

Gradient Labs builds vertical AI agents for regulated financial services, betting that full autonomy, not co-pilots, delivers safer and more compliant outcomes. Founded in London in 2023 by Dimitri Masin, Neal Lathia, and Danai Antoniou, early employees of the challenger bank Monzo, the company grew revenue nine hundred percent in a year and now supports more than thirty two million end users. It raised an extended twenty six million dollar Series A led by Octopus Ventures and CommerzVentures, bringing total funding near forty three million dollars, and counts Wise, Monzo, Zego, Pockit, Current, Stash, and Rho among its customers.

Its flagship agent, Otto, is a procedure following agent trained on a company's specific products and processes, and it resolves complex inquiries end to end rather than deflecting them. Gradient Labs offers a suite of specialist agents for lending, disputes, and know your customer checks, plus a platform that runs the operations in between. On disputes, for example, Otto gathers and verifies the evidence, decides the outcome, escalates to a human for sign off, and files the chargeback with the card scheme. It also automates business verification, claims intake, arrears collection, and fraud interception.

The agents work as interconnected specialists that share context, memory, and tasks in a multi agent ecosystem, and Otto remembers past conversations to resolve issues like freezing a lost card or tracing a missing payment. Coverage spans multiple channels including voice, and the agent works proactively and outbound, contacting a borrower in arrears at the moment they are most likely to respond or intercepting a fraud attempt live. It delivered the first ever live autonomous customer support agent for a large regulated United Kingdom bank.

Compliance is the core of the pitch. Gradient Labs describes itself as the first compliant AI platform for the financial services industry, with financial services specific guardrails and procedures that keep the agent operating safely under regulatory scrutiny, and it escalates high stakes decisions to humans for sign off. It backs its work with a money back commitment, sharing the risk if it does not deliver what it promised. Deployed as an overlay on Intercom, Zendesk, Freshworks, or in house helpdesks through an application programming interface, Otto reports resolving ninety percent of queries with ninety eight percent accuracy.