Account-level Knowledge Files and Agent Preferences (Agent Preferences on Account Command Center)

Scale your AI Agents across every account with Agent Preferences — now available to all organizations.

Agent Preferences has graduated from limited release and is now available to all NinjaCat organizations, with a new dedicated space on the Account Command Center — front and center on the main account page.

Inside Agent Preferences, eligible users can now do two things at the account level:

• Write free-form agent instructions in an open text box (e.g., tone, reporting cadence, escalation rules).
• Upload Knowledge Files specific to that account — brand guidelines, SLA and billing rules, annual advertising goals, the KPIs that customer actually cares about, and any other context that should travel with the account.

Why this matters: it's how AI Agents scale across many accounts.

Until now, giving an agent rich, account-specific context meant loading that context directly into the agent itself — which doesn't scale once you're managing dozens or hundreds of clients. With Agent Preferences, that context lives outside the agent, attached to the account, and is pulled in only when it's relevant.

The mechanism is simple: when a user @-mentions an account in an AI Agent conversation, the agent checks that account's Agent Preferences, picks up any instructions and Knowledge Files stored there, and appends them to its working context for that request. No re-uploading, no bloated agent configs, no cross-contamination between clients.

Access respects existing account permissions — the agent only sees Preferences for accounts the user is already allowed to view. Knowledge Files support up to 30 MB each (no images).

This is a meaningful step toward making AI Agents genuinely useful at agency scale, and a capability that sets NinjaCat apart in the market.