AI Governance & Human Oversight
A one-page brief for procurement, risk, and compliance reviewers.
What Oracron does
Oracron audits freight invoices: it reads each invoice, classifies every charge, compares it to your contracted rates, and surfaces overcharges with the evidence behind them — across road, air, and sea.
How AI is used — and where the human stays in control
The AI performs detection and explanation. It does not take money-affecting action autonomously. A member of your team reviews and approves before any dispute is raised with a carrier. The AI's role is to make that decision fast and well-evidenced, not to replace it.
Why this matters for your AI Act readiness
The EU AI Act (Article 14) requires effective human oversight of high-risk AI systems, applicable from 2 August 2026. Oracron's architecture is designed around that principle:
| Reviewer asks | Oracron's answer |
|---|---|
| Can a human intervene before action? | Yes — sign-off is required before any dispute is sent. |
| Are AI decisions logged? | Yes — every AI call is recorded in a tamper-resistant audit log. |
| Can a decision be traced and explained? | Yes — each finding carries its reasoning; pipeline runs are traced end-to-end. |
| Can a decision be reversed? | Yes — nothing money-affecting executes without human approval, so it can be declined or undone before it happens. |
| Where is data processed? | EU or US data region, chosen at onboarding. |
| Is customer data used to train AI? | No model training on your invoice data. Anthropic and OpenAI are used under their no-training enterprise terms; prompt/response logs auto-delete within their retention windows. |