Behavioral-AI payment security firm Trustmi launched an AI Investigation Agent on August 5, ahead of a Black Hat USA 2026 showcase. The agent connects security, finance and business teams during payment-fraud investigations, gathering evidence across systems and delivering evidence-backed outcomes rather than summaries.
Purpose-Built, Not A Chatbot
Unlike copilots that summarize or answer prompts, Trustmi says its agents perform investigative work — correlating signals from communications, documents, vendors, financial systems and human interactions, grouping related activity into a unified incident, and building a timeline of what happened and when. Investigators can request more evidence, receive next-step recommendations, and generate reports for executives. SIEM and SOC integrations deliver findings back into existing security workflows.
Part Of A Growing Agent Portfolio
The Investigation Agent joins Trustmi's AP Diligence Agent and Vendor Validation Agent, all built on the company's Behavioral AI platform for high-risk financial workflows. "Fraud investigations should not depend on teams manually piecing together evidence across disconnected systems," said co-founder and CEO Shai Gabay. Black Hat demos will cover payment fraud, vendor impersonation, business email compromise and social engineering.
Agentic Security Wave
Trustmi's launch lands amid a rush of agentic security debuts — see Astelia at Black Hat 2026, Tanium's Atlas Agentic AI and Zenity's $125M Series C to secure the era of 1B agents — as enterprises look to defend against a wave of AI-enabled B2B fraud.
Reporting based on Trustmi's press release and coverage by Security Boulevard, VMblog and SecurityBrief.