Behavioral intelligence security company DTEX Systems Inc. today introduced an expanded AI Risk Management product that reads the intent behind how employees and autonomous artificial intelligence agents use generative AI tools across the enterprise.

The release targets a gap DTEX argues most security tools have yet to close. As copilots, generative AI applications and AI agents gain access to corporate data, systems and workflows, conventional tools can log what those systems do but cannot judge whether the activity matches intent. DTEX is pitching AI Risk Management as a way to read that intent across both people and machines.

The product is designed to discover sanctioned and unsanctioned AI use across users, endpoints and workflows, covering browser, application and embedded AI activity. It flags shadow AI and embedded copilots in real time, builds inventories of approved tools and classifies the risk of unknown ones. It also monitors prompts, responses and data movement at a granular level to catch leakage of source code, intellectual property and other sensitive data.

A central claim is the ability to tell human and AI-driven actions apart. DTEX says the product delivers visibility into what it calls Computer Use AI, tracking what an agent was instructed to do, how it carried out tasks and the lineage of actions taken across systems. By correlating prompts, behavioral baselines and agent actions over time, the company says it can separate routine experimentation from risky or malicious behavior and detect agent-driven data exfiltration before it becomes a breach.

In one early deployment, DTEX said it identified an autonomous agent exposing sensitive data while operating within its intended workflow and permissions, surfacing the risk before it triggered an incident.

“AI agents are rapidly becoming operational actors inside the enterprise, with the ability to access data, interact with systems and take autonomous action,” said Chief Executive Marshall Heilman. “Many security solutions can monitor AI activity, but they still can’t determine whether that behavior aligns with intent or introduces risk.”

Alongside the monitoring tools, DTEX is adding two autonomous security agents to act on what the platform finds. Triage Guardian uses a multi-agent approach to automate investigation workflows, gather evidence and run independent reviewer agents that validate findings to cut false positives. Threat Hunter lets analysts launch investigations in natural language, then correlates findings and surfaces unknown threats on its own.

Both draw on DTEX i³ research, including work with MITRE and Five Eyes defense partners. The company said early deployments saved more than 40 hours per month per analyst.

Founded in 2000, DTEX raised $50 million from Alphabet Inc.’s growth fund CapitalG in 2024 and unifies insider risk management, data loss prevention, user behavior analytics and user activity monitoring on a single platform.

AI Risk Management is available in private preview, with broader availability expected next quarter.

Image: DTEX