BRING Your Own Agent is described as a new wave of personal AI assistants that employees connect to corporate systems, effectively acting as autonomous agents that operate in the employee’s name. The piece frames this as an identity-related challenge first and foremost: agents can wield corporate access, own or impersonate identities, and may persist beyond an employee’s tenure.
Rather than simply blocking these tools, the article argues for enabling safe usage by making identity the control plane and governing how agents access company systems.
The article explains how personal agents typically enter enterprise environments: via OAuth consent to Google Workspace, Microsoft 365 or Slack; through a plain text workflow that connects accounts; by signing in with saved passwords in the agent’s browser; using API keys or personal access tokens; or by running on the employee’s laptop (for example, OpenClaw or Muse).
It surveys several agents, including OpenAI’s Dots, xAI’s GrokBot, Meta’s Muse, and Instinct, noting that most deployments remain client-side and outside standard admin consoles. Governance guidance centres on discovering connected agents (shadow AI inventory), attributing each agent to a human owner, right‑sizing the agent’s access, and retiring the agent when the owner leaves, with tokens and sessions cleaned up.
The piece highlights the challenge of distinguishing employee versus agent actions and the need for evolving controls, such as monitoring agent‑to‑agent communications and implementing knowledge compartmentalisation during offboarding.