DTEX has introduced Deep Researcher, a new AI security agent designed to help security teams identify emerging human and AI risks before they escalate into security alerts.

The company said traditional detection tools focus on identifying known threats, while Deep Researcher is intended to uncover risks hidden within routine activity that has not been flagged by existing controls or reviewed by analysts.

Using natural language prompts, analysts can ask questions such as whether specific user behavior is normal across users, alerts, activity patterns and timelines. Rather than limiting investigations to previously tagged activity, Deep Researcher examines the broader behavioral record to determine whether available evidence supports or dismisses a potential risk.

"Every security program has a class of emerging risk where the information exists, but decision makers are completely unaware of it," said Rajan Koo, CTO of DTEX. "Deep Researcher empowers analysts to ask early, exploratory questions to validate if a risk exists."

According to DTEX, the new agent builds on the company's Agentic Defenders portfolio, which was introduced as part of its AI Risk Management initiative.

The company said Deep Researcher is powered by DTEX's lightweight digital forensic audit trail, giving the AI access to contextual information surrounding user actions, including events leading up to and following an activity as well as comparisons with similar behavior across an organization. DTEX said this broader context enables the AI to identify relationships within data more quickly than security teams could manually assemble, helping expose shadow activity that might otherwise remain undetected.

DTEX said Deep Researcher can help organizations identify risks that have not generated alerts or matched existing detections, uncover connections across users, endpoints, data and AI activity, extend the capabilities of security teams, expose shadow activity beyond traditional shadow AI and shadow IT, and produce investigation findings with evidence-backed reports for executives.

The company said the technology addresses a growing challenge for modern security teams that must sift through large volumes of alerts, logs and telemetry while prioritizing events that may indicate insider risk. DTEX noted that many insider incidents develop gradually through behavioral patterns rather than appearing malicious in isolation.

"Faster investigation is an easy claim. The better question is about the data the agent is using to reason," Koo said. "That depth is what can surface true risk nobody knows to ask about, with evidence they can act on."

Deep Researcher joins the company's existing Agentic Defenders lineup, which includes Risk Assistant for interactive chat and guided investigations, Triage Guardian for autonomous prioritization and analysis of known risks, and Threat Hunter for structured investigations based on previously tagged activity. Unlike those tools, Deep Researcher analyzes the broader untagged behavioral audit trail to identify emerging risks before they escalate.

DTEX said Deep Researcher is currently available in private preview for select customers.