Command

Command is the chat interface you use to run your workspace — the same interface, in two tiers.

Two tiers, one interface

• Customer tier (Command Center, inside the app) — a curated set of capabilities: create and change playbooks, maintain knowledge, check status, grant approvals, start and pause campaigns.
• Operator tier — the same interface with full tool access for deeper interventions, used by the cegtec team during operations.

Both tiers speak the same nine terms. What you say in chat translates 1:1 into actions on your workspace, knowledge, tables, playbooks, workflows and integrations — nothing Command does is invisible in the control room.

What you can say

Examples: “Set up a playbook for tax advisors in Bavaria”, “Show me all replies from last week with a positive signal”, “Set the ICP threshold to 70”, “Import this list and propose a structure”, “Why was this lead disqualified?” Command understands intent, not just commands — it asks follow-up questions when something is missing (budget, target region, and so on).

The approval flow

Command distinguishes proposal from execution. Structure-changing or paid actions (a new playbook, an import, a live send) are created by Command as a proposal with visible scope and a cost cap — not as an immediate action. Only your approval turns it into a running Run; see Approvals for the full rule on when approval is mandatory. Pure read access (status, evaluation, explanation) runs without approval.

Command is adapter-agnostic: it knows “email outreach” or “CRM”, never the specific vendor behind it — switching the integration changes nothing about how you phrase things to Command.