Sumsub has launched a Model Context Protocol integration enabling AI agents such as Claude and ChatGPT to configure compliance workflows.
The UK-based platform has introduced a Model Context Protocol (MCP) integration alongside a new suite of AI agent skills, allowing AI agents to access daily operations as well as the platform's configuration layer.
With the new agentic experience, an AI agent can take a compliance policy document and generate a fully configured Sumsub environment from it. A multi-page PDF containing country-specific risk brackets, weighted scoring tables, and conditional logic can be translated into platform settings such as verification levels, risk questionnaires, and onboarding workflows, directly within a customer's dashboard. Sumsub states that a setup process, which previously took days, can now be completed within minutes.
The company describes the launch as a shift in how compliance setups are built. Configuring a verification platform has typically required manual work from solution architects or technical teams, interpreting AML policies and building onboarding workflows by hand.
New capabilities for compliance teams
According to the official press release, the new capabilities cover three areas. Compliance teams can upload an AML policy and ask an AI agent to configure their Sumsub environment, with the agent reading the document and building the relevant settings. AI agents can also handle technical integration, writing the code needed to embed Sumsub verification as a mandatory step within an onboarding flow. In addition, agents can support day-to-day compliance tasks, including reviewing applicants, running analytics, generating verification links, and responding to regulatory changes.
Andrew Novoselsky, Chief Product Officer at Sumsub, said that setting up and updating a compliance workflow has always demanded considerable manual effort, particularly when regulations change. Novoselsky added that connecting an AI agent directly to the platform's configuration layer allows a team to hand over its AML policy and have the environment built automatically, calling it a new category of capability.
The integration is described as model agnostic, designed to work with AI agents generally rather than a single provider. Sumsub has published an open source set of agent skills on GitHub that can be installed with a single terminal command. At the same time, the MCP integration builds on the company's broader AI strategy, which already includes Summy, an AI Copilot for compliance and fraud teams within the platform.
Access to the MCP integration is governed by separate permissions to allow granular control over data, and sensitive actions are carried out in an isolated sandbox, with configuration changes subject to human review and approval. The integration is available now, with Sumsub stating it is the first verification platform to be officially listed on the ChatGPT Apps platform, while discussions with additional large language model providers are ongoing. Full documentation and the agent skills are publicly available through Sumsub's developer resources.
