
Variance builds AI agents that investigate fraud, sanctions and compliance cases for regulated institutions.
Variance's agents are built around an investigative loop rather than a single detection model, so they can gather context across internal systems, registries, sanctions lists and open-source sources before reaching a conclusion. Every step is logged with its evidence, which makes the output defensible for audit and regulatory review.
The platform ships a marketplace of first-party tools, including multi-list sanctions screening, adverse-media monitoring across many languages, corporate-registry coverage in dozens of jurisdictions and document or identity verification. Customers can point agents at their own procedures and integrate internal data such as their Snowflake warehouse.
Source: variance.com
Financial institutions face growing alert volumes and regulatory expectations for consistent, explainable investigations, and much of the evidence-gathering work is still handled manually by analysts. Vendors that can automate the investigative step rather than only the detection step address that bottleneck directly.
Variance positions itself for institutions regulated by banking and securities authorities in the United States, where sanctions screening and anti-money-laundering reviews are mandatory. Its stated focus is expanding agent-driven investigations across financial-crime and compliance teams.
Source: variance.com
Variance sells investigative AI agents that carry compliance and fraud cases from alert to decision for banks, fintechs, marketplaces and other regulated institutions. Agents follow a customer's written policies and procedures, gather evidence and leave a cited, auditable trail for each case.
The offering spans onboarding and screening workflows such as sanctions and PEP screening, business verification and beneficial-ownership tracing, transaction-alert investigation and customer due diligence. First-party tools and external sources are wired into the same workflow so a case can be assembled and explained end to end.
Source: variance.com