[BC Idea]: Add Support for Vision-Enabled AI Models in Business Central

BC Idea Link

https://experience.dynamics.com/ideas/idea/?ideaid=1c68ff7a-e328-f011-9d47-7c1e52f90348

Description

Description

We need to implement support for vision-enabled AI models in Business Central through the Azure OpenAI service. This would allow users to send images along with text prompts to AI models and receive detailed responses about the image content.

Why is this needed?

Vision capabilities will enable numerous valuable business scenarios:

• Product image analysis for catalog management
• Document and receipt processing with contextual understanding
• Quality control through image analysis
• etc.

This builds upon existing text-based Azure OpenAI capabilities, creating a more versatile AI integration that can work with both textual and visual data.

Requirements

• Allow sending images to AI models via multiple methods: URLs, direct uploads, Media fields, MediaSets, etc.
• Support detail level settings for image analysis (auto, low, high)
• Handle multiple images in a single request (up to 10 images per request as per Azure's limitations)
• Proper error handling for unsupported file types
• Documentation and examples for integration

Reference

Implementation should follow the official Microsoft documentation for GPT with vision:
Azure AI Services: Use vision-enabled chat models

I will provide the implementation for this BC Idea

• I will provide the implementation for this BC Idea
  Internal work item: AB#578341

Start with the existing text-based Azure OpenAI capabilities and the linked Microsoft documentation for vision-enabled chat models. Confirm how URLs, uploads, Media fields, and MediaSets should map to requests, including detail levels, image limits, and unsupported types. Done means documented examples and working support for the listed image inputs and error cases.