Trustmi launches AI investigation agent for B2B fraud

Trustmi has launched an AI Investigation Agent for B2B fraud detection, expanding its AI platform with a specialist agent.

The tool is designed to investigate suspicious activity across payment-related workflows and bring together evidence from multiple systems for security and finance teams. It uses automated investigative workflows to connect activity across communications, documents, vendor records, financial systems and human interactions.

The launch reflects a broader push by cybersecurity and finance software providers to address fraud that spans multiple business systems at once. In many cases, finance teams, security teams and banking partners work from separate data sources, slowing efforts to determine whether a payment request, vendor change or account update is legitimate.

Trustmi's AI Investigation Agent is intended to address that problem by gathering and correlating evidence, linking related activity into a single incident record and building a timeline of events. Users can also ask the system questions, request more evidence, receive suggested next steps and prepare reports for senior management.

Cross-team focus

Trustmi is positioning the product around collaboration between departments that often investigate fraud separately. The agent can bring security, finance and business teams into a shared investigation process instead of leaving each group to work from partial information.

The system also integrates with SIEM and SOC tools already used by security teams, allowing alerts and findings to flow into existing operational processes. This is intended to reduce the need for finance teams to rely on technical specialists to interpret security data, while helping security teams understand payment activity that may sit outside their usual remit.

The product enters a market where fraud schemes increasingly rely on social engineering, vendor impersonation and business email compromise rather than purely technical attacks. Such incidents can appear credible because they exploit normal business approvals and established payment procedures, making them harder to detect through standard controls alone.

According to Trustmi, the AI Investigation Agent extends its existing platform beyond fraud detection alerts into a more active investigation model. The system can request new information, gather additional evidence and tailor findings to different stakeholders as an inquiry develops.

"Fraud investigations should not depend on teams manually piecing together evidence across disconnected systems," said Shai Gabay, Co-founder and Chief Executive Officer of Trustmi.

"The AI Investigation Agent helps connect the right people, evidence and context in real time, enabling finance and security teams to collaborate more efficiently and ask the right questions as the investigation unfolds. By reducing manual effort and accelerating the path to an evidence-backed conclusion, organizations can act faster to stop fraud before it results in financial loss."

Wider platform

The agent joins other specialist products in Trustmi's portfolio, including AP Diligence Agent and Vendor Validation Agent. The broader platform is focused on financial workflows that carry elevated fraud risk, particularly around accounts payable, vendor onboarding and payment approvals.

Trustmi describes itself as a behavioral AI payment security provider. It says it protects more than USD $240 billion in payments each year and has helped prevent more than USD $1 billion in fraud and USD $5 billion in payment errors.

That emphasis on payment fraud comes as companies face pressure to tighten controls without slowing day-to-day operations. Businesses often run approval and payment processes across email, enterprise resource planning systems, supplier databases and document repositories, leaving investigators to reconstruct events after money has already moved.

By seeking to create a single view of those workflows, Trustmi aims to give finance and security teams a common evidence base during an incident. It argues that faster reconstruction of events can improve the odds of halting fraudulent transfers or containing the impact before losses escalate.

Trustmi was founded in 2021 and is headquartered in New York City. Its backers include Cyberstarts and Insight Partners.

The new AI Investigation Agent is part of a growing portfolio of specialized agents built for high-risk financial workflows.