Someone asked me whether Microsoft runs a secret agent called Scout instead of Copilot. Scout is real, it was on the Build stage, and the name is the interesting part. It is not marketing. The name follows the architecture.
Copilot is a promise Scout does not keep
The word Copilot is a contract with the user. A copilot waits. You ask, it drafts, you edit. Microsoft spent three years teaching 400 million seats to expect exactly that. Scout does something categorically different. It runs on a schedule you set, watches your inbox and your Teams messages, executes shell commands, drives a browser through Playwright, and finishes work while you are in a meeting. Ship that under the Copilot name and you corrupt both promises: people who expected a chat box get an agent that moved their calendar, and people who wanted the agent go hunting for it in a sidebar. Microsoft named the new category Autopilots.
Identity is the real reason
Scout gets its own Entra identity. It acts on your behalf as a distinct principal, with permissions that can be narrower than yours, inside a zero-trust runtime where the agent's own container is treated as untrusted and identity, tokens, and policy sit outside it. An agent with its own identity cannot be a feature of your Copilot session. That is not a branding preference. It is an architectural fact expressing itself as a name.
The moment the agent is a separate principal in your directory, it needs a separate noun in the product line, a separate row in the audit log, and a separate answer for your CISO. If it stayed a ghost inside the user's session, audit logs would blur human and machine, least privilege would be aspirational, and incident response would be archaeology.
Two more reasons round it out. Packaging: Scout needs two license SKUs, a Frontier gate, and an attestation, and you cannot upsell a feature, only a category. Blast radius: an always-on agent with shell access will eventually do something expensive, and Microsoft would rather the headline name a preview product with 3,000 users than the brand on the entire productivity suite.
The lesson
If you find yourself arguing about what to name an agent, check whether you are actually arguing about an architecture. Name the category when the identity model, the permission surface, and the failure modes change, not before and not to be clever. For the full story, including how ClawPilot went from 100 to 3,000 users in a week and why that curve is the only pre-launch eval that means anything, read the full post.