Topic starter
02/06/2026 6:55 pm
TL;DR: 63% of organisations experienced at least one AI-related security incident in the past year, incidents rose by more than 50% year over year, and more than half of deployed AI agents are not actively monitored or secured, according to Cyera’s cited SACR technoscope. The control problem is no longer perimeter protection; it is continuous oversight of behaviour, data in motion, and intent across agentic systems.
Editorial analysis by NHI Mgmt Group, based on content published by Cyera: “SACR Names Cyera an Innovator in 2026 UADP Technoscope”.
Key questions
Q: How should security teams govern AI agents that can change actions at runtime?
A: Security teams should govern runtime AI by correlating identity, data, and intent before trusting an action path.
Q: Why do legacy DLP controls fail in agentic workflows?
A: Legacy DLP fails when the risky event is the agent’s decision chain rather than a file leaving the environment.
Q: What should teams look for when AI agents are not being actively monitored?
A: Teams should look for agent inventories that do not map to an owner, an approved data scope, and an observable action boundary.
Practitioner guidance
- Map AI oversight to a unified control path Inventory where identity, data classification, and enforcement are still managed separately, then define the runtime decision point that must join them for agentic systems.
- Review agent monitoring coverage Identify deployed AI agents that are not actively monitored or secured, and tie each one to an accountable owner, data scope, and approved action boundary.
- Test controls against prompt injection paths Validate whether instruction manipulation can change tool use, data access, or downstream actions without triggering your current detection and response model.
Bottom line: AI and data security are converging because agentic systems force identity, data, and enforcement into one runtime control problem.
Explore further
View Full Forum → | NHI Foundation Course → | Our Services → | Read the full analysis →
02/06/2026 7:10 pm
Unified agentic defense is an architectural response to control fragmentation. The article’s real signal is that AI security and data security are converging because the old split between identity, data, and enforcement no longer matches how agentic systems operate. When systems can reason over context and act in runtime, separate point tools create blind spots between policy, data classification, and execution. The implication is that governance will increasingly be judged by whether it closes those gaps across the full decision path.
A question worth separating out:
A: Organisations should treat them as closely linked governance functions because AI safety controls and cybersecurity controls overlap in practice. The same programme should cover testing, vulnerability management, model use, and responsible deployment. Splitting them too far apart creates gaps between policy, engineering, and operational oversight, which makes it harder to define due diligence and consistent accountability.
👉 Read our full editorial: AI and data security are converging around agentic defense platforms