GeoComply Launches MCP Server for AI Fraud Investigations

GeoComply Solutions Inc. announced the launch of GeoComply MCP on September 29, 2026, a product the company said gives fraud and compliance teams access to its risk signals, compliance statuses and location intelligence through the AI assistants those teams already use.

The Vancouver-datelined announcement presents the product as a replacement for an investigation workflow in which transaction history sat in one tool, device and network links in a second, and location and compliance checks in a third. GeoComply said that work runs in a single conversation: the analyst asks a question in plain language, and the assistant retrieves the relevant data and tools and runs the analysis. Each verdict returns with the signals behind it attached, while the decision itself remains with the analyst.

According to the company’s product page, GeoComply MCP is a remote server built on the Model Context Protocol, an open standard from Anthropic that lets an AI assistant use an outside system as a tool. It works with Claude, ChatGPT, Copilot and other assistants that support the standard, and a customer’s own agent can reach the same endpoint machine to machine.

What GeoComply MCP Does

The release sets out the product’s core capabilities. Analysts can ask about a user, a device or a network of connected accounts and receive an assembled answer with the evidence attached, without writing SQL or opening a data-team ticket. Playbooks built on GeoComply’s investigation expertise are intended to give new analysts an experienced investigator’s standard from day one, the company said, and teams can adapt the playbooks to their own processes. Compliance staff can pull pass rates and non-compliant transaction breakdowns across brands, markets and products on request rather than waiting on a scheduled report. Every scored verdict shows the signals behind it, which GeoComply said lets analysts defend a call before regulators, in chargeback disputes or in internal review.

“Fraud and compliance teams are full of sharp analysts, and their time should go into making the right decisions, quickly and accurately,” said Steven Vo, Chief Technology Officer at GeoComply. “Pulling the evidence together is exactly the work a machine should be doing, and doing it consistently and reliably is what GeoComply’s compliance-grade risk, device and network signals provides.”

The product page details the tools behind the interface. Investigation tools return a scored fraud verdict for any user or device, with the transaction history, the linked-account network and the signals behind the score delivered in one pass. Network-mapping tools follow shared devices and connected accounts out from a single starting point, and a bulk lookup accepts up to 500 usernames or device IDs at once, returning block status and connection counts. Transaction search covers a year of history for a user or device, with a full profile available for any single transaction.

Further tools handle deep spoofing analysis, geographic searches of transactions, users and devices inside a polygon, radius or street address, geofence analysis, and block and allow-list browsing with pass-rate and error breakdowns by operator, solution, source or region. Troubleshooting tools convert a failed rule into a documented cause with a fix and an escalation flag, explain error codes, and search the documentation behind them. Identity-verification lookups run alongside geolocation investigations for customers holding the IDComply entitlement, and Edge analytics require per-organization onboarding.

Access Paths

GeoComply said teams can reach the same intelligence through whichever route fits their workflow. Any AI assistant that supports the Model Context Protocol can connect, without tying the team to a single vendor. GeoComply can stand up an investigation bot in workspaces such as Slack, with nothing for analysts to install. Engineering teams can point their own agent or model directly at the GeoComply MCP, and the company said the same scoping and audit controls apply to every call. Inside its own back office, GeoSentry, the company’s built-in AI analyst, handles the same questions within GeoComply Explorer, which customers already sign in to. On the customer’s side, the product page said, an organization admin adds the connector in the AI platform and signs in once with single sign-on, with no software deployment required.

Security, Audit and Rollout Terms

Every request is scoped server-side to what the user and the organization are permissioned for, and every action is logged with who ran what and when, according to the release. The product page states that data isolation is enforced in two places: each call is scoped to the operators an organization is entitled to, and the underlying platform enforces that scope again using the organization’s own credential. Analysts inherit the permissions they already hold through single sign-on, per-organization tool allowlists cap the set of tools any caller can reach, and the MCP reads and reports rather than acting, with case actions still handled in the customer’s existing case-management tools. GeoComply maintains SOC 2 Type II and ISO 27001 and is aligned with GDPR and CCPA.

“Speed only matters if you can trust the signals underneath,” Vo said. “Whether a team is investigating ATO syndicates, payment fraud, or money laundering, the answer is only as good as the data behind it.” Vo described GeoComply as fast becoming the signal layer those teams reach for first, and said the MCP makes it easier to put the signals to work.

The MCP is an AI Feature under GeoComply’s Artificial Intelligence Features Addendum, the release noted. Under those terms, AI-generated output can be incomplete or inaccurate and is intended to support a trained reviewer rather than replace one, and customers are responsible for their own human-led review of any output before relying on it.

GeoComply said it is rolling out GeoComply MCP to customers, directing them to their GeoComply contact or the product page for more information.