When Security Awareness Training No Longer Covers Who's Working

November 19, 2026
Shannon Lewis

The Expanding Perimeter No One Authorized

Organizations rolled out AI tools to accelerate decisions and reduce manual workloads. Then security teams discovered employees had already deployed agents they never approved. These tools now hold credentials, access data, and execute tasks without oversight.

The gap is not technical. It is structural. Security frameworks were built to govern people, not autonomous systems. That misalignment creates exposure in permissions, visibility, and response protocols.

This tension surfaced during KnowBe4's Lagos executive event on human and AI risk. The discussion centered on a question security leaders are now forced to answer: how do you govern a workforce that includes both humans and AI agents?

Why This Matters Now

Adoption of AI-driven automation has outpaced security policy updates. Tools are introduced through trial accounts, departmental budgets, and individual subscriptions. By the time security teams gain visibility, agents have already been granted permissions modeled after human roles.

The risk is not hypothetical. Attackers are already testing AI-generated phishing campaigns that adapt based on recipient behavior. These campaigns exploit gaps in awareness training that assume human recipients and static threat models. When agents interact with compromised systems or execute commands from manipulated inputs, traditional monitoring fails to flag the activity as anomalous.

Security awareness programs were designed for static user bases. They assume employees receive training, absorb protocols, and apply judgment. Autonomous agents do not process training modules. They execute based on permissions and integrations. That difference forces a shift from awareness to governance.

KnowBe4's event in Lagos positioned this as a workforce risk management problem. Organizations must now secure two populations: humans who need behavioral training and agents that require access controls, monitoring, and audit trails.

Three Strategic Gaps Exposed

Employees Deployed Agents Before Security Teams Gained Visibility

Shadow AI is not introduced through formal procurement. It arrives through free trials, departmental purchases, and integrations that bypass IT approval. By the time security teams learn which tools are in use, agents have already been granted access to email, file systems, and external APIs.

  • Security teams lack inventory of active agents across the organization
  • Agents are integrated into workflows without documentation or audit trails
  • Permissions are granted without understanding the scope of agent actions
  • Decommissioned tools leave orphaned credentials and access rights

Permissions Were Copied from Human Roles Without Adjusting for Agent Behavior

Organizations often grant agents the same permissions as the employees who deployed them. That model assumes agents will exercise judgment and escalate ambiguous decisions. Agents do not. They execute commands based on integrations and inputs, even when those inputs are malicious or malformed.

  • Agents inherit excessive permissions because they mirror human roles
  • Access reviews do not account for non-human identities
  • Agents operate continuously, expanding the window of exposure
  • Credential rotation policies do not extend to agent API keys and tokens

Security Frameworks Were Not Built to Monitor Autonomous Systems

Traditional security monitoring flags anomalies in user behavior: unusual login times, unexpected file access, or geographic inconsistencies. Agents operate outside those patterns. They work continuously, access data at high volumes, and execute tasks across time zones. Existing tools do not distinguish between legitimate automation and compromised agents.

  • Monitoring systems do not differentiate between human and agent activity
  • Incident response playbooks assume human actors, not autonomous systems
  • Threat intelligence feeds focus on human-targeted attacks, not agent manipulation
  • Security teams lack visibility into which agents are active, dormant, or misconfigured

The Strategic Shift Required

Security teams must transition from managing awareness to governing a hybrid workforce. That means treating agents as first-class identities with distinct risk profiles. Awareness training remains necessary for employees, but it must be paired with visibility, access controls, and monitoring for autonomous systems.

This shift requires integrating agent governance into existing security frameworks. Organizations need inventory management for active agents, permission models that reflect agent behavior, and monitoring systems that flag anomalies in automation patterns.

The goal is not to eliminate AI adoption. It is to govern it in a way that maintains visibility and control. That requires policy updates, tooling adjustments, and collaboration between security, IT, and business units.

  • Establish centralized inventory of all active agents across the organization
  • Implement permission models designed for autonomous systems, not human roles
  • Deploy monitoring that flags agent anomalies distinct from user behavior
  • Build incident response protocols that account for compromised or misconfigured agents
  • Integrate agent governance into existing security awareness and compliance programs

How Agent Risk Manager Addresses This

KnowBe4's Agent Risk Manager is designed to provide visibility and control over the hybrid workforce. It combines awareness training for employees with governance tools for AI agents, positioning both under a unified workforce risk management framework.

  • Employees Deployed Agents Before Security Teams Gained Visibility: The platform offers real-time visibility into which agents are active across the organization, including shadow AI introduced without IT approval. It provides inventory management and audit trails for agent deployment and integration.
  • Permissions Were Copied from Human Roles Without Adjusting for Agent Behavior: Agent Risk Manager includes permission modeling designed for autonomous systems, enabling security teams to review and adjust access based on agent function rather than mirroring human roles. It supports automated threat detection for misconfigured or overprivileged agents.
  • Security Frameworks Were Not Built to Monitor Autonomous Systems: The platform delivers monitoring and control tools that distinguish between human and agent activity, flagging anomalies specific to automation patterns. It integrates with existing security awareness programs to provide a unified approach to workforce risk management.

Who This Is For

  • IT security leaders managing both human and AI-driven workflows
  • CISOs responsible for governing shadow AI and autonomous systems
  • Security engineers tasked with monitoring non-human identities
  • Organizations adopting AI tools without clear visibility or governance frameworks

If your team is struggling to govern shadow AI and agent permissions, see how KnowBe4 Agent Risk Manager provides visibility and control. Visit https://knowbe4.optrics.com/ai-agent-risk-manager.html

FAQ

What is shadow AI?
Shadow AI refers to AI tools and agents deployed by employees without formal IT approval or oversight. These systems often operate outside security frameworks, creating visibility and governance gaps.

Why do agents require different permissions than employees?
Agents execute tasks continuously and do not exercise judgment. Permissions modeled after human roles often grant excessive access because they assume discretion and escalation. Agents need access controls that reflect their behavior and scope.

How does Agent Risk Manager integrate with existing security programs?
The platform extends security awareness training to include agent governance. It provides inventory management, permission modeling, and monitoring tools that work alongside existing compliance and training initiatives, creating a unified workforce risk management approach.

What types of organizations benefit most from agent governance tools?
Organizations with distributed teams, high AI adoption rates, and complex permission structures benefit most. This includes enterprises managing multilingual workforces, multinational subsidiaries, and environments where automation is already embedded in daily operations.


Optrics Logo white shadow
Optrics is an engineering firm with certified IT staff specializing in network-specific software and hardware solutions.

Contact Information

6810 - 104 Street NW
Edmonton, AB, T6H 2L6
Canada
Google Plus Code GG32+VP
Direct Dial: 780.430.6240
Toll Free: 877.430.6240
Fax: 780.432.5630
Copyright 2026 © Optrics Inc. all rights reserved. 
linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram