Cimento emerges from stealth to secure the one thing no firewall can protect

Cimento, an AI-native human risk management startup, has emerged from stealth with $3M in pre-seed funding. The platform builds real-time behavioral risk profiles for every employee by ingesting signals from existing security tools (VPN, CrowdStrike, DLP, cloud data lakes) and running a Bayesian model to generate per-user risk scores. Its standout feature is multi-turn phishing simulation — running iterative, multi-channel attack campaigns (email, SMS, voice, WhatsApp) against high-risk user segments rather than one-off phishing emails. When risk thresholds are crossed, the platform can automatically revoke access via Okta without manual intervention. The company is also building toward AI agent risk management, currently mapping agent risk to the human operator. Early customer Together AI highlights the gap in existing tools like KnowBe4 and Proofpoint when it comes to AI agent-derived identities and implied trust.