Integration with AI

Path: Integration with AI · tier official · system copilot · narrative reviewed (checked by Opus)

Overview

This area groups four ways to bring AI into Business Central. Developer tools for Copilot cover generative AI features written in AL, including the System.AI module, the PromptDialog page type and Azure OpenAI setup. The machine learning APIs cover the Forecasting (Time Series) and Prediction APIs.

The Model Context Protocol pages describe the Business Central MCP Server and the AL MCP Server, and how to connect them from hosts such as VS Code and Copilot Studio. The agents (preview) pages explain how to design, create, configure, run and code custom agents with the AI development toolkit in sandboxes.

Start with the Copilot developer tools if you are building generative AI features in AL. Choose the machine learning APIs for forecasting or prediction tasks. Go to the MCP pages to expose Business Central to external AI hosts. Use the agents pages for custom agents that work on tasks.

Key points

• Copilot developer tools: what counts as a Copilot extension, the System.AI module, the PromptDialog page type, Azure OpenAI setup, Business Central AI resources, Help preparation and a transparency note.
• Machine learning APIs: the Forecasting (Time Series) API and the Prediction API, plus a transparency note on capabilities, limits and responsible AI.
• Business Central MCP Server: overview, configuration, API tools, dynamic tool mode and data query tools.
• MCP connection options: VS Code, Copilot Studio and non-Microsoft hosts, with access-control guidance.
• AL MCP Server: supports AL development tools.
• Agents (preview): design, create, configure, run, iterate and code custom agents with the AI development toolkit in sandboxes.
• Agent topics include instructions, permissions and profiles, tasks, attachments, visibility, export/import and the Tasks AL API.

Subtopics

• Integrate AI using developer tools for Copilot (24 pages)
• Develop AI with machine learning APIs (3 pages)
• Integrate with Model Context Protocol (6 pages)
• Designing and coding agents (preview) (23 pages)

Videos, posts and code changes

Linked by a Haiku matcher with a grounding quote from the item’s summary (link/topics.ts); machine-generated.

• #10205 Deprecate Text Completion API in AI SDK (work item 621038) (code change): “Text Completion API in the Azure OpenAI AI module is marked obsolete”
• Why not start improving Business Central APIs definitions for supporting AI? (community post): “Business Central APIs lack AI-ready metadata that LLMs need to reliably call functions”
• What’s New: Business Central AI Resources (2025 release wave 1) (video): “ai resources; copilot; azure openai; content safety; data privacy”

Business Central pages and reports

The pages, reports and queries this hub’s Learn pages name (ms.search.form, API reference), joined to the code by exact object id (data/index/docs-objects.json); a page shows the table it runs on.

• Page 8350 “MCP Config List” · captioned “Model Context Protocol (MCP) Server Configurations” · via Integrate with Model Context Protocol
• Page 8351 “MCP Config Card” · captioned “Model Context Protocol (MCP) Server Configuration” · via Integrate with Model Context Protocol
• Page 8359 “MCP Config Warning List” · captioned “MCP Configuration Warnings” · on Table 8352 “MCP Config Warning” · via Integrate with Model Context Protocol

Source: Microsoft Learn (CC BY 4.0). Descriptions are Learn’s own.