Retrieval on Tech Foundations

Conceptual Alignment

Mon, 01 Jan 0001 00:00:00 +0000

Conceptual Alignment and Contract Discipline in RAG Systems

A recurring problem in RAG system design is conceptual misalignment. Different parts of the system often use the same term for different artifacts, or collapse distinct artifacts into a single object.

Model

Mon, 01 Jan 0001 00:00:00 +0000

Model of Context-Augmented Task Systems

This ontology and models defines a systems-level view of context-augmented task execution. Its purpose is to separate concepts that are often collapsed in retrieval-augmented systems: the user’s observable request, the task the system infers from that request, the context required to complete the task, and the source-directed queries or actions used to obtain that context.

Model Representation Design

Mon, 01 Jan 0001 00:00:00 +0000

Task need	Useful representation
Find exact term or identifier	lexical index, keyword field, exact-match metadata
Find semantically related policy or explanation	dense vector representation, hybrid search
Find a person, product, account, or project	entity index, normalized metadata
Answer from tables	table objects, extracted rows, schema-aware search
Traverse relationships	graph representation
Verify authority	source metadata, ownership, approval status, policy hierarchy
Check freshness	timestamps, version fields, update history
Cite evidence	passage-level chunks with provenance
Summarize large documents	section summaries, document outline, hierarchical chunks

A stronger framing would be:

Reframing

Mon, 01 Jan 0001 00:00:00 +0000

Beyond RAG: Task-to-Context Alignment

Goal: We should stop treating RAG as the architecture and instead treat it as one pattern within task-to-context systems.