Abnormal Security

Abnormal AI, listed here under its former name Abnormal Security, is a behavioral security platform. Instead of matching known signatures, it models how each person and machine identity inside an organization normally operates, learning their work, access and relationships, then evaluates messages and actions against that baseline and flags meaningful deviations. The platform covers three areas. Email security stops business email compromise, advanced phishing, vendor fraud and account takeover, with products for inbound filtering, an AI security mailbox that acts as a co worker for the operations team, just in time phishing coaching, misdirected email detection and data loss rules. Identity security handles account takeover protection, identity threat detection, infiltration prevention and posture management. AI security governs the organization's own AI adoption, with cloud security, employee guardrails, agent security and a security workbench. Underneath sits an identity trust graph that maps human, non human and cloud identities, fed by signals from Microsoft 365, Google Workspace, Entra, Ping Identity, Slack, Azure, GCP and Copilot. This is enterprise software for security teams rather than a self serve tool. Pricing is not published. There is no pricing page on the site, and the only route offered is a demo request, so terms are quoted per organization after a sales conversation. The company was founded in 2018 by a team whose earlier work was in advertising technology, where they built models that analyzed large datasets and made anomaly based decisions in milliseconds. The company describes applying that same behavioral approach to cyber defense, and it publishes benchmark comparisons of its detection against general purpose language models.