Knowledge Layer

Retrieval, RAG & Knowledge

Explain how enterprise AI systems locate, rank, validate and authorize knowledge.

Why This Matters

Enterprise answers are only as trustworthy as the retrieval path that produced them. Ranking, provenance, and access control are architecture problems, not prompt tricks.

Architecture Position

Retrieval is the Knowledge Layer: it feeds context assembly and must respect Trust Layer authorization before a model ever sees a chunk. See it on the architecture map.

Start Here

Core Concepts

Retrieval Foundations

  • Semantic search
  • Keyword search
  • Embeddings
  • Vector databases

Retrieval Strategies

  • BM25
  • Hybrid retrieval
  • Reranking
  • Metadata filtering
  • Query rewriting
  • Query decomposition

RAG

  • Basic RAG
  • Advanced RAG
  • Agentic RAG
  • Multi-stage retrieval
  • GraphRAG

Data Pipelines

  • Ingestion
  • Chunking
  • Indexing
  • Refresh
  • Document processing

Enterprise Retrieval

  • Access-aware retrieval
  • Permission-aware RAG
  • Data residency
  • Provenance
  • Citations

Evaluation

  • Retrieval precision
  • Recall
  • Groundedness
  • RAG evaluation

Advanced

  • Knowledge graphs
  • Deterministic retrieval
  • Retrieval planning

Advanced

Advanced material

Deeper guides for this domain are on the way.

Coming Soon

Architecture Patterns

Applied patterns for this domain live in the pattern library.

Browse architecture patterns

Case Studies

Enterprise walkthroughs that apply this topic in a full system.

Browse case studies

Related Topics

Roadmaps

This topic appears in the Enterprise AI Architect path.

Follow the Enterprise AI Architect roadmap
Go deeper

Learn this as part of the Enterprise AI Architecture course.

Need help applying this architecture to a real system?