Forrester’s Vision Report | Securing Intent: A New Security Domain For Agentic AI

Forrester’s Vision Report

Securing Intent: A New Security Domain For Agentic AI

Agentic AI moved intent beyond the user and the prompt. Agents now choose their own path to a goal, and reasoning loops, memory, and tool calls each become a surface to secure. In this report, Forrester defines agentic intent as its own security domain, maps the 5 ways security programs fail without it, and lays out the capabilities and playbooks to govern it.

Forrester names Cyera as one of its example data security vendors for detecting and protecting sensitive data flows at the agent layer.

INSIDE THE REPORT

• Forrester's working definition of AI agentic intent, and why intent is no longer user  or prompt-centric
• The five failure modes of security programs that ignore intent, from misclassifying risk to treating agents as service accounts, and the fix for each
• The difference between task adherence and goal adherence, and the new class of risk between them: technically compliant behavior that undermines the human's actual purpose
• The Intent Impact Matrix: four quadrants that separate accidental harm from purposeful harm, so the response matches the motive
• The 5 layers of intent (maker, organizational, role-based, user, agent) and the separate evidence each requires before you classify a behavior
• 8 core capabilities for intent security, including how DSPM and DLP must meet the runtime agent 

Information in Forrester publications is based on Forrester’s efforts to compile and analyze the best resources reasonably available to Forrester at any given time. Opinions reflect judgment at the time and are subject to change.  This report is part of a broader collection of Forrester resources, including interactive models, frameworks, tools, data, and access to analyst guidance.