This session is a practical guide to building private, governed AI within Microsoft Dynamics 365 Business Central using Microsoft Foundry as a centralized control plane for model management, orchestration, and cost visibility.

Microsoft Foundry is presented as the “brain” of the architecture, responsible for model selection, evaluation, secure execution, and coordination of AI-driven workflows.
The session walks through the end-to-end integration path: connecting Business Central with Foundry, enabling secure inference from AL, orchestrating tool usage and retrieval, and applying production-ready patterns such as request/response contracts, throttling, fallbacks, validation strategies, and telemetry.

It also introduces Model Context Protocol (MCP) and proxy-based approaches that enable AI to perform structured actions directly within ERP workflows, including inventory intelligence, data validation, and contextual recommendations.

Finally, the session highlights the shift from assistive AI to agent-driven workflows, where AI moves beyond chat-based interactions into goal-oriented execution within business processes.
Attendees will leave with practical patterns, reference architecture, and a checklist to implement their first Foundry-powered capability in ERP.