According to a recent LinkedIn post from Astrix Security, Software Analyst Cyber Research has released a report arguing that securing AI agents is fundamentally an identity challenge rather than a model issue. The post notes that the report introduces an Agentic Identity Access Platforms framework, addressing agent discovery, intent-based authorization, just-in-time access, and runtime enforcement.
Claim 30% Off TipRanks
- Unlock hedge fund-level data and powerful investing tools for smarter, sharper decisions
- Discover top-performing stock ideas and upgrade to a portfolio of market leaders with Smart Investor Picks
The post further indicates that the report concludes legacy IAM and PAM systems are ill-suited to AI agents, which operate with dynamic credentials, broad intents, and limited human oversight. Highlighted risks include unmanaged “shadow” agents, over-privileged credentials, hardcoded secrets in MCP servers, and a lack of runtime visibility into agent activity.
As shared in the post, Astrix has produced its own write-up on where its offerings may fit within this emerging framework and what security teams should prioritize. For investors, this emphasis on AI agent identity and access control suggests a potential expansion area in cloud and application security, positioning Astrix to benefit if organizations increase spending to address these newly surfaced risks.
The focus on gaps in traditional IAM and PAM could underscore a growing demand for specialized tools that secure machine identities and autonomous agents as AI adoption scales. If enterprises validate these concerns with budget allocations, vendors aligned with the AIAP approach, including Astrix, may gain a competitive edge in a nascent but potentially significant segment of the cybersecurity market.