Sumsub lets AI agents build compliance setups through new MCP integration

Sumsub has rolled out a Model Context Protocol (MCP) integration that lets AI agents configure an entire compliance environment from an AML policy document. For PSPs and high-risk merchants, the practical point is simple: the gap between written policy and live onboarding settings just got a lot smaller.

1. AI agents, including Claude, ChatGPT and other leading models, can now take a compliance policy document and translate it into a fully operational Sumsub setup. That includes verification levels, risk questionnaires and onboarding workflows, all configured directly in the customer’s dashboard.
2. The company says this replaces a manual process that previously required technical teams to interpret complex AML policies, including country-specific risk brackets, weighted scoring tables and conditional logic. In Sumsub’s framing, a task that once took days can now be completed in minutes.
3. The release covers three main use cases: teams can upload AML policy or regulatory requirements and ask an AI agent to configure the Sumsub environment; agents can write the code needed to embed Sumsub into a client application and build verification into the onboarding flow; and they can support ongoing compliance work by reviewing applicants, running analytics, generating verification links and adapting to regulatory changes.
4. Sumsub says the MCP integration is model-agnostic and works with any leading AI agent. It has also published an open-source collection of agent skills on GitHub, installable via a single terminal command.
5. Access is controlled through separate permissions for granular data control, and sensitive actions run in an isolated sandbox environment. Any configuration changes still have to be reviewed and approved by a human before they take effect.

Sumsub already positions itself as a full-cycle verification platform for fraud-free, scalable compliance, and it also offers Summy, an AI copilot for compliance and fraud teams inside the platform. The new MCP layer pushes that idea one step closer to “policy in, setup out” — which is exactly the kind of workflow PSPs and merchants will care about when compliance rules change faster than the ops team can update the form builder.