Question

Your engineering team is building an Azure AI Foundry agent that assists field technicians. Technicians upload photos of equipment nameplates and hand-drawn wiring diagrams, and the agent must interpret the visual content, reason about it, and generate step-by-step textual repair guidance in natural language. The solution must handle both image understanding and complex text generation within a single deployed model to minimize integration overhead. Which model selection best fits these requirements?

Accepted Answer

Deploy a multimodal LLM (such as GPT-4o) that accepts both image and text inputs and produces text output

Answer

Deploy a text-only small language model (SLM) and add a separate OCR-only preprocessing step

Answer

Deploy an embedding model to convert the images into vectors for similarity search

Answer

Deploy a text-to-image generation model to render repair diagrams