Many finance functions are experimenting with AI, but few have unlocked meaningful ROI. The reason isn’t a lack of technology; it’s the absence of a clear operating model that connects data, people, and platforms to business outcomes. This discussion explores the following themes:

· Defining “real ROI”: What measurable value from AI looks like in finance operations

· The accuracy paradox: Setting realistic expectations for AI versus human performance

· Organisational readiness: New roles, skills, and operating rhythms for AI-augmented finance

· Data as fuel: Embedding quality, governance, and accessibility for AI acceleration

· Platform integration: ERP, IBP, and analytics modernisation as foundations for success

· Human-in-the-loop: Designing oversight frameworks that balance risk and speed

· Implementation lessons: How to move from experimentation to scalable adoption

· Model governance & assurance: Establishing finance-grade validation, testing, and auditability for AI models, agents, and automated decision flows

· Enterprise risk & control alignment: Embedding AI into existing risk frameworks, ensuring controls, escalation paths, and accountability structures are updated for new operational, cyber, and compliance risks.