Blog

Latest Insights

Thoughts on AI agent governance, security, and building trust into autonomous systems.

March 30, 2026·Alex Floyd

Enterprise Data Labels: Enforce Purview, Google DLP, and AWS Macie Labels on AI Agents

AI agents don't respect data classification labels. Your enterprise spent months configuring Microsoft Purview, Google DLP, or AWS Macie to tag every document with sensitivity levels. But when an AI agent processes that document, those labels are invisible. Clevername now reads enterprise sensitivity labels on incoming content and blocks agents that don't have clearance. No reclassification. No new taxonomy. Your labels, your rules.

March 29, 2026·Clevername Team

Inside Clevername's Security Architecture: Internal Evaluation Results

We subjected Clevername's security architecture to two internal evaluations: a 10-phase runtime suite (632 tests) and an 8-phase governance lifecycle suite (184 tests, 100% pass rate). Combined: 816 test cases covering content scanning, credential detection, review approval, drift detection, enforcement parity, compliance evidence, and more. Here are the results.

March 22, 2026·Clevername Team

How to Route AI Agent Traffic Through Clevername From Your Existing Firewall

Your firewall sees AI API traffic but can't inspect what agents are doing with it. Here's how to forward AI-bound requests through Clevername's Guard API for scanning, drift detection, and governance — without replacing your existing security stack.

March 22, 2026·Clevername Team

Why Your AI Agent Framework Doesn't Protect You

LangChain, CrewAI, and OpenAI Agents SDK give you powerful tooling. They do not give you security. The OWASP Agentic Top 10 shows exactly where the gaps are — and what to do about them.

March 22, 2026·Clevername Team

The CISO's Blind Spot: What Happens Between Agent Approval and Agent Action

Identity controls who. Scanners check what. Nobody enforces the approved mandate continuously. That gap between approval and action is where 88% of AI agent security incidents originate.

March 22, 2026·Clevername Team

Detecting API Keys in AI Agent Prompts: A Problem Nobody's Talking About

Users paste API keys into chat prompts every day. The LLM sees them. The training pipeline might ingest them. This is a real, measurable risk — and we built a system to stop it.

March 20, 2026·Alex Floyd

Why Agent Governance Is the Missing Layer in Your AI Stack

Identity platforms ask who an agent is. Scanners inspect what it said. But nobody is enforcing what it's actually allowed to do — continuously, from approval to every action.

March 18, 2026·Alex Floyd

Introducing the Agent Review: Human Sign-Off for Every Agent

No agent should go live without human review. The Agent Review brings quorum-based approval, security questionnaires, and guardrail profiles to every agent deployment.