Aembit IAM for Agentic AI now supports Microsoft Copilot Studio. Read More →

Get an Aembit Tenant
TALK TO AN ENGINEER
Blog

Tag: AI Workloads

AI chat interface connected to enterprise data systems, illustrating how sensitive information can be exposed through prompts, agents, and model access.

What Causes AI Data Leakage and Tips for Staying Protected

As AI moves from chat windows to enterprise systems, data leakage becomes an identity and access problem.
Dan Kaplan
Jun 2026
|
9 min read
Two software engineers at a desk with dual monitors displaying code, with the blog title overlay: What Anthropic WIF Gets Right – And What It Doesn't Solve

Anthropic Workload Identity Federation: What It Gets Right – and What It Still Doesn’t Solve

Eliminating static API keys is real progress – but securing one credential surface is not the same as governing workload access at scale.
Apurva Davé
Jun 2026
|
8 min read
Software engineer seated at a desk in a modern office, reviewing responses from an AI assistant on a laptop while working across multiple monitors in a natural, candid workplace setting.

AIMS: A Model for AI Agent Identity

An early IETF draft hints at how identity infrastructure may evolve once autonomous software starts acting inside enterprise environments.
Kevin Sapp
May 2026
|
4 min read

Gartner’s Workload IAM Architecture Is a Big Step Forward for AI Agent Security

The global research and advisory firm is pushing the industry toward a more practical model for securing AI agents and non-human access.
Kevin Sapp
May 2026
|
8 min read
MCP Server Access

Auditing MCP Server Access and Usage

AI agents exchange sensitive contexts across MCP servers in seconds. Without context-aware auditing, you can’t trace who accessed what.
Ashur Kanoon
May 2026
|
6 min read
Context-Based Access Control for MCP Servers: Why Static Rules Fail

Context-Based Access Control for MCP Servers: Why Static Rules Fail

Static access rules fail in dynamic MCP environments. Context-based access control evaluates identity, context and resources in real time.
Dan Kaplan
Apr 2026
|
6 min read
Dark horizontal diagram showing a “Delete Resource” action enabled by a redacted access token, leading to a production database dissolving and a “resource permanently deleted” warning.

How a Long-Lived API Credential Let an AI Agent Delete Production Data

An agent behaved like a true insider threat. Unmanaged API keys made those mistakes devastatingly consequential. Both can be true at the same time.
Dan Kaplan
Apr 2026
|
4 min read
Introducing the Aembit MCP Server.

Analyze AI Agent Access: Introducing the Aembit MCP Server

Teams can query workload identity data in plain language, investigate activity, and move faster without leaving the Aembit platform.
Ashur Kanoon
Apr 2026
|
4 min read
JIT Access for Wordkloads

Just-in-Time Access for Workloads: Eliminating Standing Privileges

Most workload credentials, the API keys, tokens and passwords that connect your services, carry “always on” access that never expires.
Apurva Davé
Apr 2026
|
6 min read
Diagram showing IAM for agentic AI, where workloads use identity providers and short-lived credentials to securely access APIs, data platforms, and services.

Aembit IAM for Agentic AI Is Now Generally Available

Built in the open with customers, now ready to run against real agent workflows.
Ashur Kanoon
Apr 2026
|
5 min read
Load More