This case study is a finalist in the SAPinsider APAC 2026 Innovation Vanguard – Customer of the Year category.

Putting an AI agent into a live finance operation requires more than a model that works. It needs a clearly defined role, trusted evidence, human decision rights and named accountability. Village Roadshow put these principles into practice when it deployed DIANA—the Duplicate Invoice Analysis and Notification Agent—in approximately eight weeks to identify suspected duplicates before payment.

DIANA now reviews the complete population of approximately 132,000 supplier invoices across A$425 million in annual AP spend, operating with 99.4% detection accuracy. Hear how Village Roadshow moved from sample-based review to population-level assurance, retained human control over every decision and established an Agent Boss, Agent Coach and executive mandate—creating a repeatable model for taking future AI agents into production across Finance and the wider enterprise.

What You’ll Learn

• What it takes to select, validate and move an AI agent into a live finance process.
• How to combine AI recommendations with human judgement and auditable controls.
• How clear ownership and ongoing coaching can support accountable AI at scale.