Insider risk teams can now see what AI agents do with company data, and stop sensitive data before it leaves
MOUNTAIN VIEW, Calif., Nov. 5, 2026 /PRNewswire/ -- Cyberhaven, the leader in data security for the agentic enterprise, today introduced Agentic Data Security (ADS), to help CISOs, data security teams, insider risk teams, and AI governance leaders say yes to AI agents without losing control of their data.
Agentic Data Security shows security teams what every AI agent does with company data, who it is acting for, where the data goes next, and stops sensitive data before it leaves. The companies leaning furthest into the agentic future, from frontier AI labs to top quant hedge funds, already rely on Cyberhaven to protect their most sensitive data. ADS is available for purchase today.
AI agents have changed how work gets done, and with them comes a new kind of insider risk. As agents move from pilots to production, security leaders across industries are asking the same question: what is an agent doing with our data, and who sent it? Agents act with the access of the people they serve, at machine speed, so a single agent can read a file, query a database, and share what it finds within one workflow. Cyberhaven Labs found that endpoint-based agentic AI app adoption doubled year over year, reaching 60% in May 2026. This rapid adoption creates novel data risks that can originate with "trusted" agents.
"AI agents now work with the access of a trusted employee," said Nishant Doshi, Cyberhaven CEO. "The first agentic data incident at most companies may be an agent doing exactly what it was asked to do, with data it should never have touched. Most companies will adopt agents faster than they can build new controls for them. Agentic Data Security gives them the confidence to deploy agents while keeping control of the data itself, not just the tooling that touches it."
Agentic Data Security is the latest module in Cyberhaven Flow, the unified data security platform that uses data lineage, AI, and autonomous protection to deliver data loss prevention (DLP), insider risk management (IRM), and data security posture management (DSPM) for humans and agents. It represents the next step in the technology strategy the company introduced in July, which brings data lineage, identity, and behavior together to protect data across every human and agentic workflow.
"One cannot govern AI-era risk without understanding what data was touched, by whom, through what workflow, and whether that movement was authorized," said Lawrence Pingree, Head of Research, SACR, a leading industry analyst firm. "Cyberhaven's data lineage gives security teams that foundation for people and AI agents together, in one platform."
Keep the Agents, Keep Control of the Data
Cyberhaven's Agentic Data Security gives security teams four controls over every agent:
- Find every agent. Discover every AI tool and agent running across endpoints, any cloud or browser, including unsanctioned apps, and score each one for risk.
- See what it does. Capture prompts, tool calls, file reads, reasoning traces, and responses, along with the data involved in each step. Because they are documented like any other content, every agent action is open to inspection.
- Know who sent what. Tie each action to the person or system behind it, with people and agents in one view.
- Stop data before it leaves. Apply existing policies for people to agentic actions to identify risky behavior, and contain sensitive data before it leaves, without blocking teams from adopting AI.
"The agentic enterprise will be defined by how well companies control the data agents touch, not by how many they deploy. Agents change the math on data exfiltration: they move faster, at scale, with trusted access," said Dan Walsh, CISO, Datavant. "The winners will say yes to AI without watching their data walk out the door, and that requires security that follows the data everywhere it goes. That's the direction Cyberhaven is set to lead."
Cyberhaven traces the full lifecycle of data, adapting protection as context changes. Agentic Data Security brings that to agents with complete visibility into workflows through data lineage, one unified architecture for both humans and agents, and protection that takes action in real time through enforcement across endpoints, browsers, and the cloud. ADS applies data lineage and a unified policy engine to agents, without adding another tool, console, or set of policies.
"Data security in the agentic era will be defined by understanding AI data lineage," said Manish Kalia, Chief Product Officer at Cyberhaven. "Our proprietary AI models help customers not only classify their data accurately but also understand its flow as AI moves and transforms it, so Cyberhaven can protect data at rest and in motion. Because agent tool calls and reasoning traces are content Cyberhaven can already inspect, the policies customers use to manage insider risk now apply to agents too."
Go deeper on securing the agentic enterprise:
- Join the live executive briefing, "Double Trouble: Taking Control of Human & Agentic Insider Risk" on December 3, 11:30 AM CST.
- Read the blog from Manish Kalia: Agentic Data Security: Control What AI Apps and Agents Do With Data
- Read "Securing the agentic enterprise with data lineage," MIT Technology Review Insights, October 2026.
- Watch the Agentic Data Security Walkthrough: https://vimeo.com/1235907613?fl=pl&fe=sh
About Cyberhaven
Cyberhaven is the leader in Data Security for the Agentic Enterprise, trusted by the world's most innovative companies to protect their most sensitive data. Cyberhaven traces the full lifecycle of your data, adapting protection as context changes. Cyberhaven protects workflows, not just data, by connecting lineage, identity, and behavior to understand how data is created, copied, fragmented, and shared across workflows. Operating in real time where humans and agents act on data, Cyberhaven helps stop risk before it becomes loss. AI changed work. Cyberhaven protects it. Learn more at cyberhaven.com.
SOURCE Cyberhaven
Share this article