Proofpoint, Inc., the cybersecurity and compliance company, has announced Proofpoint AI Security, a security solution that combines intent-based detection, multi-surface control points and a comprehensive implementation framework to secure how humans and AI agents use AI across the enterprise.
Based on the industry-first Agent Integrity Framework, which defines how an AI agent operates with integrity, Proofpoint is introducing a five-phase maturity model for implementation, from initial discovery through runtime enforcement.
As organisations rapidly deploy autonomous AI agents to browse the web, access internal systems, send emails, execute code, and orchestrate workflows, risks such as agentic privilege escalation and zero-click prompt injection attacks are becoming real-world threats. A single AI request can trigger dozens of autonomous actions across multiple systems, often at machine speed and without human oversight.
Existing security tools can see traffic and permissions, but they don’t evaluate whether AI behaviour aligns with the original user intent. Research from Acuvity, recently acquired by Proofpoint, indicates that 70% of organisations lack optimised AI governance, and 50% reported expectations of AI-related data loss within 12 months, highlighting the urgent need to secure agentic AI environments.
“AI is now embedded in how work gets done, and security must evolve with it,” said Sumit Dhawan, CEO of Proofpoint. “Humans and AI agents share similar risks: both can be manipulated and both can take actions that diverge from their intended purpose, yet traditional security was never designed to validate intent. Proofpoint is uniquely positioned as a unified cybersecurity platform built to protect people, defend data, and govern AI agents together, providing continuous, intent-based verification that behaviour aligns with policy and intent in the agentic workspace.”
Intent-based Detection Models across all AI Interactions
Traditional security tools lack visibility into the semantic content of AI interactions, where risk often resides. They can see traffic, identities, or permissions, but they cannot determine whether an AI’s actions are appropriate in context or aligned to the intent of the user or agent.
Proofpoint AI Security closes this gap by applying intent-based detection models, continuously evaluating whether AI behaviour, initiated by a human or an autonomous agent, aligns with the original request, defined policies, and intended purpose. By analysing the full semantic context of AI interactions, the solution flags misaligned or high-risk actions in real time, before damage occurs, such as non-compliant communication or data loss.
Protecting Against the Full Spectrum of Autonomous AI Risks
Delivered through a unified architecture for the agentic workspace, Proofpoint AI Security operates across the surfaces where AI is used – including endpoints, browser extensions, and MCP connections – giving organisations the visibility and control needed over AI usage and risks. This is particularly critical in developer environments, where agent-connected coding assistants, plugins and MCP-integrated tools are accelerating adoption and increasing the need for visibility and policy enforcement.
Through these control points, organisations can:
- Discover sanctioned and unsanctioned AI tools such as OpenClaw, Ollama, ChatGPT, and MCP servers used by humans and agents
- Observe prompts, responses, and data flows during AI tool usage
- Apply access controls and guardrails on AI usage
- Implement runtime inspection and enforce policies during live AI interactions
Structured Path to Implementation with Agent Integrity Framework and Market-First AI Maturity Model
To provide enterprises with a structured roadmap to safely govern AI, Proofpoint is also introducing the Agent Integrity Framework, a comprehensive guide that defines what it means for an AI agent to operate with integrity and provides a five-phase maturity model for implementation, from initial discovery through runtime enforcement.
The framework defines Agent Integrity as the assurance that an AI agent operates within the boundaries of its intended purpose, authorised permissions, and expected behaviour—across every interaction, tool call, and data access. It outlines five pillars: Intent Alignment, Identity and Attribution, Behavioural Consistency, Auditability, and Operational Transparency, to help customers operationalise AI governance without needing to overhaul their existing security architecture.
“Humans are expected to operate with integrity when using business systems, and AI agents must be held to the same standard,” said Ryan Kalember, executive vice president of cybersecurity strategy at Proofpoint. “Agent Integrity means ensuring that AI agents act within the boundaries of their intended purpose, authorised permissions, and expected behaviour across every interaction, tool call, and data access. With Proofpoint AI Security and the Agent Integrity Framework, we can provide a clear blueprint to help enterprises comprehensively address the full spectrum of risks that emerge when AI agents operate autonomously across enterprise systems.”
