Skip to content

Prompt Security (now part of SentinelOne)

Runtime GenAI and agentic-AI security company preventing AI-related data leakage and protecting AI agents, acquired by SentinelOne for ~$250M.

Visit Website ↗
68/100Incremental Innovator

Overview

Prompt Security focused on securing generative and agentic AI at runtime — monitoring and controlling interactions between users, employees and AI tools/agents to prevent sensitive data leakage into external AI services and to protect deployed AI agents from manipulation. Its positioning is closer to a GenAI-focused extension of data-loss-prevention (DLP) than to model-layer security.

SentinelOne completed its acquisition of Prompt Security on September 5, 2025, for a total consideration of roughly $133.6M in cash plus stock and assumed options (reported deal value around $250M in some coverage), as part of SentinelOne’s strategy to extend its Singularity Platform to cover GenAI and agentic AI use cases end to end.

Innovation Matrix Assessment

Innovation Velocity 7/10

Built runtime GenAI protection covering both data-leakage prevention and agent security, and was acquired within roughly two years of the generative AI security wave beginning.

Operational Value 7/10

Runtime interception of AI interactions gives security teams a practical DLP-style control point for GenAI usage, an immediately actionable capability.

Market Momentum 8/10

Acquired by SentinelOne, a major public endpoint-security vendor, for a deal SentinelOne itself disclosed at roughly $133.6M cash plus equity — a strong, primary-source-confirmed validation event.

Category Disruption 6/10

Extends the well-established DLP control pattern to GenAI traffic; valuable but conceptually adjacent to existing data-security tooling rather than an entirely new control type.

Real-World Efficacy 5/10

No independent, third-party efficacy validation of its detection or blocking accuracy was found.

Enduring Relevance 8/10

Preventing sensitive-data leakage into external AI tools and securing agent behavior are durable needs as GenAI and agentic AI usage keeps expanding inside enterprises.

Why CISOs Should Care

Gives security and data-loss-prevention teams visibility and control over what data employees and AI agents send to external and internal GenAI tools, closing a gap traditional DLP wasn't built to cover.

What Makes It Different

Applies runtime interception and policy enforcement specifically to AI/agent traffic rather than requiring security teams to retrofit generic web/DLP proxies for GenAI use cases.

The Matrix Verdict

68/100 — INCREMENTAL INNOVATOR

Strong Performer on momentum, thanks to a SentinelOne-confirmed acquisition price, but its conceptual approach (DLP extended to AI) is scored more conservatively on disruption than category-native approaches like model-layer red-teaming.

Editorial Note: Claims vs. Verified Findings

Deal terms are corroborated by SentinelOne's own press release and SEC-related filings. Founding year and pre-acquisition headcount were not found in this research pass and are marked Undisclosed rather than estimated from general knowledge.

Sources