Articles, whitepapers, and videos on putting AI agents to work in the enterprise - safely and at scale.
AI security is not another pillar in the Cyber Defense Matrix. It is the entire spectrum. A single enterprise AI agent - built by one person in Finance on a Tuesday afternoon, connected to your ERP, your reporting system, and your communication tools - simultaneously implicates every row of the matrix in a single interaction. You cannot govern this by assigning it to a cell.
Roee Salomon
Most enterprise AI governance programs are notarization services. They witness the policy. They do not verify the control. A notarization does not verify the underlying facts - it witnesses the signature. The OpenAI incident illustrates what happens when an agent operates without governance that verifies controls - not just documents them.
Amir Ofek
Human-in-the-loop governance is the most widely cited answer to the enterprise AI agent risk problem. It is also the answer that fails most reliably in production. Not because the concept is wrong. The instinct to keep a human accountable for consequential AI actions is correct. The problem is that “human-in-the-loop” as an implementation strategy breaks down in three distinct ways before most enterprises realize it has broken - and the replacement architecture is rarely in place when it does.
Chen Pipek
Showing 16 - 18 of 40 articles
PDFA routine billing complaint. Four lines of routing text. One AI agent with access to your CRM. That was all it took. aizome researchers documented a complete indirect prompt injection attack chain using only a public Salesforce Web-to-Case support form - no credentials, no exploit, no Salesforce bug. One support ticket caused a privileged AI agent workflow to export customer records and deliver them to an external mailbox. The attacker’s only contribution was text in a support form.
Download PDF
PDFNew research from SACR defining the control layer that decides whether a live AI agent action should complete, not whether the agent had access an hour ago. Covers the three-layer architecture, the Agentic Control Depth model that sets ACD 4 as the production bar for agents touching sensitive data, and a vendor landscape of more than thirty providers. aizome is named a pure-play ARISE vendor.
Download PDF
PDFSACR names aizome an Emerging Leader in agentic identity and access. The report covers why traditional IAM was not built for agents, and lays out a four-phase model for securing them: discover, authorize, broker, and watch. aizome is one of six vendors profiled, and one that spans all four phases.
Download PDFSee aizome in action. Your employees are already building AI agents - connected to your most sensitive systems, operating under inherited identities, with no one watching what they’re actually doing. This video shows what changes when aizome is in your environment. Every agent discovered. Every identity governed. Every action validated before it executes. Every outcome accountable. Watch to see what’s already running in your environment - and what it looks like when it’s governed.
Your AI agents don’t sleep. They don’t take breaks. And if nobody is watching their intent - not just their access - they will reach further than they should. In this episode, aizome CEO Amir Ofek joins Hacker Valley Media to talk about the autonomous workforce already running inside enterprise environments - ungoverned, unmonitored, and unknown. A conversation about why knowing what an agent is permitted to do is not the same as knowing what it is actually doing right now. Listen here - https://hackervalley.com/e/do-you-know-what-your-agent-is-doing-at-3am-with-amir-ofek
NHI is evolving. Or is it being replaced? In this fireside chat, SACR analyst Lawrence Pingree and Amir Ofek make the strongest possible case for why extending NHI frameworks to cover AI agents is not the right strategy. The argument: organizations treating enterprise AI agent governance as “NHI with extra steps” are building programs that will fail in production. ARISE - Agentic Runtime Identity Security Enforcement - is not an extension of NHI. It is what comes after it.
Amir Ofek
Showing 1 - 3 of 4 videos
PDFWhere does aizome sit against the emerging agentic AI risk framework? This control mapping shows exactly - 25 control objectives, six domains, three coverage levels: Native (aizome delivers directly), Contributes (aizome provides critical enforcement and telemetry), and Out of scope (organizational processes aizome deliberately doesn’t claim to own). No overclaiming. No gaps hidden in fine print.
Download PDF
PDFaizome is the Enterprise AI Agent Control Platform. This document outlines every capability across six domains - what each feature does, how it works, and the business value it delivers.
Download PDF
PDFA datasheet on how Aizome makes every enterprise AI agent accountable, with identity, intent and control for the agents your workforce is already building. See what the platform delivers, how it fits the stack you already run, and how to get visibility in minutes.
Download the datasheet (PDF)Showing 7 - 9 of 9 datasheets
PDFIn a world of ten thousand faces, not every identity is what it seems. In this chapter of The Indigo Guardian, Ito confronts the dangers of false identities, wandering agents, and invisible threats. But true protection isn’t about stopping everything - it’s about knowing what belongs, recognizing what doesn’t, and establishing boundaries that preserve trust. Through Japanese-inspired storytelling, aizome explores a fundamental principle of agentic AI security: You cannot protect what you cannot see, and you cannot govern what you cannot identify. The first battle isn’t access. It’s identity.
Read the comic
PDFFour samurai. One clan. One crosses a boundary he was never meant to cross. In this chapter of The Indigo Guardian, discover how even the most trusted agents can drift beyond their intended purpose - and why identity, visibility, and governance are essential in the age of agentic AI. Through the art of Japanese-inspired storytelling, aizome explores a critical security challenge: How do you protect what you cannot see? Every agent needs an identity. Every identity needs a guardian.
Read the comic
PDFIn Japan, indigo protected what it touched by becoming part of it, not by standing between. Mamoru, miru, shiru - to protect, to see, to know. The story behind the name, in a single page.
Read the comicOccasional, substance-first writing on identity, intent, and putting AI agents to work across the enterprise. No noise.
