SEON has added a Model Context Protocol server, Network Detection, AI Chart Builder and an AI Playbook to its fraud prevention and anti-money laundering platform. The additions also let customers connect external AI tools to SEON data.
The expansion targets fraud and AML teams that already use generative AI tools but struggle to bring investigation data, customer profiles and risk signals into those systems securely and efficiently. SEON's new MCP server opens its data layer to third-party AI environments, including ChatGPT, Claude, Gemini and Microsoft Copilot.
Analysts can use the connection to access more than 900 real-time risk signals spanning identity, device, behavioural, AML and IP data. The server uses the open MCP standard, allowing customers to switch between AI tools without rebuilding the connection.
The move comes as AI use in fraud operations becomes more common, while integration across fraud and AML systems remains patchy. In its 2026 Fraud & AML Leaders Report focused on APAC, SEON found that 97% of organisations had integrated AI into daily workflows and 96% were confident in its reliability, but 77% still found it difficult to obtain a unified view across fraud and AML systems.
Only 51% reported having fully integrated workflows. That gap matters for teams using AI in investigations, where fragmented data can limit analysis and create operational and security risks when staff move information manually between systems.
Open access
Tamas Kadar, Chief Executive Officer and Co-Founder of SEON, said the company wanted to give risk teams more freedom in how they use AI.
"The software world is moving toward a headless model, where teams don't need to live inside a vendor's dashboard to get full control over data and functionality," Kadar said.
"Our job is to be the best command center for fraud, risk and compliance intelligence. We're giving analysts the freedom to use whichever AI tools work best for them."
Some customers want to run more of their risk work through agent-based systems, while others prefer to stay within SEON's own interface. The new products are designed to support both approaches.
Network Detection builds on analysis tools SEON released earlier, including similarity ranking and a network graph. It scans the previous two months of transactions across devices, email addresses, phone numbers and IP addresses to identify clusters that may appear suspicious only when viewed together.
The feature is intended to help analysts identify coordinated fraud rings and money laundering networks earlier in an investigation. AI Chart Builder, meanwhile, lets users ask questions in natural language and generate visualisations based on live SEON data rather than spreadsheet exports or separate dashboard projects.
Customer use
One customer cited by SEON said the MCP connection had changed how its analysts handled investigations.
"The SEON MCP integration has fundamentally changed how our risk analysts operate," said Eric Taylor, Manager of Trust and Safety at TurboTenant.
"Before, they had to manually pull data across multiple systems to piece together what happened. Now, we pull a user's entire platform journey and all of SEON's risk signal context directly into Claude, and AI connects the dots on complex fraud patterns without us doing that assembly. It's opened up OSINT capabilities that wouldn't have been possible before."
SEON said the new products join AI-assisted rule creation, scoring insights, AML screening analysis, automated case summaries and regulatory report generation introduced over the past year. The latest additions are accompanied by an AI Playbook for Risk and Compliance Teams, which includes guidance on connecting AI tools to SEON and setting up investigation workflows.
The playbook also includes pre-built agentic skills such as a fraud analyst daily briefing and a decline spot-check. Those are compatible with the MCP server and can be deployed immediately by customers.
Another customer said the broader shift in online risk is forcing companies to rethink the systems they use to monitor fraud and identity checks.
"The next generation of fraud and KYC challenges won't look like the last one. AI agents will interact with our marketplaces as customers, and AI agents will be used to impersonate and exploit our customers as well. Our team needs an intelligence foundation that's ready for both," said Mostafa Hassanin, CISO at SMG Marketplace.
"SEON opening its data layer to any AI we want to use is exactly the kind of architectural decision that fits where the market is going."