Use advisors and AI permissions

Hire project advisors, activate their method in a conversation, and understand what they can never do.

Advisors are methods, not collaborators with accounts

The Advisors dimension presents official series and film advisors. Hiring an advisor attaches a read-only methodology to the project. A conversation can activate the relevant advisor so the AI reasons with that lens.

An advisor does not become a project member, does not receive CRDT write access, and cannot act independently. The AI cannot hire, dismiss, or select an advisor for you. Those choices stay in the user interface.

Hire deliberately

Review the advisor's purpose, then hire it for the current project. The current product enforces a single-advisor workflow; dismiss the existing advisor before hiring another. Membership requirements, if any, are enforced by the visible product guard and backend rather than by the advisor itself.

Activate the lens in conversation

Use an advisor when you want a particular diagnostic method, not as a decorative title. Ask the AI to apply the method to a bounded artifact: an outline sequence, a character arc, a scene, or a Cinema reel. You can still instruct it to analyze only and wait before editing.

Project permissions win

OWNER, ADMIN, and EDITOR can write project data; VIEWER is read-only. The AI tool set follows that boundary. A collaborator's craft identity—director, writer, producer, and other roles—is presentation metadata and does not grant editing rights.

Worldview has an additional content-specific gate: even an editor's AI cannot modify it without an explicit write request in the latest message. This protects durable project rules from conversational inference.