Knowledge & Data · Software component
Adaptive Retrieval Controller
Software componentKnowledge & DataKnowledge & DataVariation point (abstract)arc:AdaptiveRetrievalController
An abstract retrieval-gating component that decides per query whether to retrieve external knowledge or answer from the model's parametric memory, and which retrieval strategy to apply.
Responsibility. Decides whether and how to retrieve for each query.
Also known as: Adaptive retrieval, Retrieve-vs-generate decision
Variants
| Variant | When to choose |
|---|---|
| Confidence-Gated Retrieval Controller | Choose when a simple gate suffices and the model's self-assessed certainty is reasonably calibrated; combine with pattern rules and periodic audits to catch miscalibration. |
| Query-Type Retrieval Router | Choose when query types differ in their optimal retrieval strategy and more sophisticated routing than a single confidence gate is warranted. |
Relationships
is configured by structural
is invoked by dependency
emits telemetry to dynamic
routes to dynamic
Design guidance
- SHOULD skip retrieval for queries the model answers correctly from parametric knowledge.
- SHOULD force retrieval for domain-specific patterns (e.g., product or policy queries) regardless of model confidence.
- SHOULD fall back to parametric generation when retrieval fails or returns no results.
- SHOULD evolve routing rules and thresholds from logged per-query-type performance rather than keeping them static.
Quantitative guidance
As stated by the sources; verify before use.
- 20-40% of production queries target general knowledge answerable without retrieval (Ch6.6).
Classification
- Patterns
- Adaptive retrieval
- Quality attributes
- Cost efficiencyPerformance efficiency (ISO/IEC 25010)Reliability (ISO/IEC 25010 | NIST AI RMF: valid and reliable)
- Risks mitigated
- Unnecessary retrieval cost and latency for general-knowledge queriesNoise from tangentially related retrieved chunksStale parametric answers for product-specific queries
Sources
- Ch6.6: T. Nguyen, "Query Decomposition and Adaptive Retrieval," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 6.6. ISBN: 9798244538229.