AI Copilot
Grounded, cited and always in your voice.
Overview
The AI Copilot is the platform's ambient assistant. It drafts customer updates, answers grounded questions, and recommends next best actions — always with citations to the source record.
Business Value
Cuts customer communication time by 70%.
Consistent, professional voice across every operator.
Every recommendation is defensible via citations.
How It Works
The Copilot is grounded in the Digital Twin repository. It never invents state — every claim is bound to a field. Actions require explicit human approval and land in the audit trail.
Workflow
- 1
Ask
Natural language question about any shipment, AOG case, KPI.
- 2
Cite
Read the source chips to verify the answer.
- 3
Approve
Confirm any recommended action explicitly.
Key Components
Best Practices
- ▸Always read the citation before approving an action.
- ▸Ask 'why' — the Copilot will explain its reasoning path.
- ▸Use the prompt library for consistency across the team.
AI Insights
Never ungrounded
If the Copilot cannot cite it, it will not claim it — a hard rule.
Frequently Asked Questions
Can the Copilot take actions on its own?›
No — every action requires explicit human approval and produces an audit event.
Which model powers it?›
The Copilot is model-agnostic; grounding and citation logic are the moat.