Sensitive actions are approval-gated and audit logged

DO
Workspace setup

Define what a profitable lead looks like

Your scoring, campaign guidance, sales capacity, and margin warnings are only as good as these operating assumptions.

Business profile

Used in campaign packages, lead scoring, and revenue reports.

Economic guardrails

These values drive qualification and margin warnings. Weak inputs create weak recommendations.

Approval gates

Money movement and customer-facing execution stay under human control.

Campaign launch
Require owner/admin approval before any live launch.
Budget increase
Block spend increases until the owner reviews evidence and risk.
First outbound sequence
Review the first message before an automated sequence begins.
Invoice creation
Validate scope and deterministic totals before issuing an invoice.
Finance export
Require approval before exporting bookkeeping data.

Scoring baseline

Current setup is strong enough to rank leads, but one risk remains.

Service area defined
Minimum job value defined
Margin floor defined
Sales capacity defined
Customer profile documented
Add service-specific minimum values before using the score for automatic routing.

What the AI may do

The operating boundary is explicit.

DraftOffers, ads, follow-ups, quote narratives
ScoreLeads, campaign health, exception risk
RecommendNext action with evidence and confidence
PrepareLaunch packages, invoices, finance exports
It may not increase spend, send bulk outreach, issue refunds, or post accounting entries without approval.