An AI workforce for financial crime compliance - agents that do the manual AML work banks would rather not, in real time.
The compliance analyst that never sleeps, never files late, and hands the regulator a receipt for every decision it makes. Founded 2024. Headquartered in New York.
There is a peculiar economics to financial crime compliance, which is that most of the work is not, strictly speaking, finding crime. It is closing alerts about things that turn out to be nothing. A payment flag that resolves to a duplicate name. A business whose ownership structure is opaque only because the paperwork is boring. Someone, somewhere, on salary, reads each one and clicks a button. Arva AI's founding idea is that a very large share of those buttons can be clicked by software - provided the software can explain itself afterward.
That last clause is the whole business. Plenty of companies will sell a bank an AI model. Far fewer will sell one that a bank's risk committee, its auditors, and eventually its regulator will tolerate touching a live compliance decision. Arva builds enterprise-grade AI agents for anti-money-laundering (AML) reviews - screening, Know Your Business and Know Your Customer (KYB/KYC) onboarding, and transaction monitoring - and it wraps them in the thing regulated buyers actually care about: explainability and an audit trail.
The company was founded in 2024 by Rhim Shah and Oli Wales, went through Y Combinator's Summer 2024 batch, and in January 2025 raised a $3 million seed round led by Gradient Ventures, Google's early-stage AI fund, with Y Combinator, Amino Capital and Olive Tree Capital along for the ride. The team is small - on the order of 15 to 20 people - and unusually specialized: financial crime, machine learning, and the specific art of getting the two to coexist inside a bank.
What Arva is selling, in the end, is not the removal of humans from compliance. It is the removal of humans from the 90% of compliance that never needed a human in the first place. The agents take the routine reviews; the analysts take the genuinely ambiguous ones. Whether you call that replacing analysts or un-burying them depends mostly on whether you are one.
Arva's agents autonomously handle millions of reviews each month. Each product targets a specific, expensive corner of the AML workflow.
Triage and resolution for sanctions, adverse media and politically exposed persons (PEP) alerts - the false-positive flood that eats analyst hours.
Customer due diligence and enhanced due diligence powered by deep web analysis, collapsing multi-day onboarding reviews into seconds.
Autonomous handling of AML transaction-monitoring alerts, taking on the alert fatigue that legacy rules engines generate.
The control platform where institutions configure Arva's agents and watch them work - oversight and audit trails built in, not bolted on.
Figures per Arva AI; approximate and self-reported.
Previously led the FinCrime product team at Revolut Business, where he saw first-hand where manual compliance work piled up. Studied Engineering at the University of Oxford. Holds the CAMS anti-money-laundering certification.
Former lead product engineer at Opvia (YC S20), with earlier full-stack work at Iventis and The Trade Desk. Studied Computer Science at the University of Cambridge.
Our CDD and EDD processes are powered by Arva - quality of analysis is consistently high.
Arva has transformed how we handle screening reviews - it resolves alerts automatically.
From the outset, Arva stood out - AI could be trusted in a regulated environment.
Launches via Y Combinator (S24) with AI agents for instant global KYB onboarding.
Closes a $3M seed round led by Gradient Ventures, with Y Combinator, Amino Capital and Olive Tree Capital participating.
Announces partnership with FairPlay to advance powerful and compliant agentic AI in financial services.
Profile compiled from public sources. Metrics are approximate and self-reported by Arva AI.
Arva AI builds enterprise-grade AI agents that do the manual grunt work of financial crime compliance - screening, KYB/KYC onboarding, and transaction monitoring - for regulated banks and fintechs. Its agents autonomously resolve the majority of alerts and reviews (roughly 91% of screening alerts, 87% of KYB/KYC assessments, and 86% of transaction-monitoring alerts), cutting onboarding times from days to seconds and compliance operations spend by up to 80%. Founded in 2024 by Rhim Shah and Oli Wales and part of Y Combinator's Summer 2024 batch, the New York-based company raised a $3M seed round led by Google's Gradient Ventures in January 2025.
Last updated: