This module develops the skill of designing multilingual NLP architectures—integrating Azure AI Translator with Azure AI Language capabilities to support global use cases—and the broader service combination design skill of selecting and assembling the right combination of capabilities for a given conversational AI or text analytics scenario. Learners extend the single-language architecture designed in Module 1 to a full five-language multilingual deployment.

This module develops the skill of designing Azure AI Vision solutions—selecting the appropriate prebuilt capabilities for image analysis, real-time video analytics, physical space monitoring, and video processing scenarios, and determining when custom vision model training is required—using a structured decision framework that prevents both over-engineering with unnecessary custom training and under-engineering with insufficient prebuilt capabilities.

This module develops the skill of designing enterprise document-processing solutions using Azure AI Document Intelligence—selecting appropriate prebuilt extraction models for standard document types, designing custom model-training pipelines for non-standard forms, and integrating extracted structured data into downstream systems. The insurance claims processing architecture is completed by connecting the document intelligence pipeline to the vision architecture from Module 1.

This module develops the skill of designing complete Azure AI Speech solution architectures—selecting between real-time and batch transcription using a structured decision framework, configuring custom speech models for domain-specific vocabulary, and designing text-to-speech solutions—with latency, scalability, and cost as the primary architectural constraints.

This module develops the skill of architecting multimodal solutions that integrate Azure AI Speech with Azure AI Language and Vision capabilities within a Microsoft Foundry project, designing the service integration architecture, orchestration pattern, and unified output design that produces a coherent multimodal solution from services that were designed independently.

This module develops the skill of designing the knowledge mining ingestion and enrichment architecture—configuring Azure AI Search indexers for large-scale document archives, designing AI enrichment skillsets that apply built-in and custom cognitive skills to extract structured insights from unstructured content, and designing knowledge stores for downstream analytics consumption. You will work through the complete ingestion architecture for a 500,000-document legal firm archive, making the built-in vs. custom skill decision for each enrichment requirement.

This module develops the skill of designing the query and retrieval architecture that makes an enriched knowledge base useful for intelligent search, including configuring semantic ranking, hybrid search pipelines, and custom scoring profiles to optimize relevance for the specific query patterns and content characteristics of a large legal document archive. Learners will explore the interaction between scoring profiles and the semantic ranker, including default behavior where scoring profiles shape initial L1 candidate selection before semantic scores re-rank the final results.

Learners produce a portfolio-ready Cognitive and Knowledge Services Architecture document for a provided enterprise scenario—integrating NLP, document intelligence, vision, speech, and knowledge mining capabilities into a unified multi-service solution architecture that mirrors the deliverable a practicing Azure AI Solutions Architect would produce for a complex, multi-capability enterprise AI engagement.