Agentic AI Reference Architecture

Observability & Evaluation · Software component

Keyword Match Scorer

Software componentObservability & EvaluationObservability & Evaluationarc:KeywordMatchScorer

A response scorer that detects required or forbidden phrases in a response and converts their capped count into a normalized score.

Responsibility. Scores presence of required or forbidden phrasing such as empathy acknowledgments.

Also known as: Key-phrase fuzzy matching in integration tests

Variant of Response Scorer abstract

When to choose. Choose for cheap detection of required/forbidden language (e.g., acknowledgment phrases, unauthorized guarantees).

alternative toevaluatesreadsspecializesis invoked byLLM Judge: alternative toLLM JudgeReAct Agent Controller: evaluatesReAct Agent ControllerEvaluation Dataset: readsEvaluation DatasetResponse Scorer: specializesResponse ScorerAgent Test Runner: is invoked byAgent Test Runner
Direct neighbourhood (hover for relationship types)

Relationships

is invoked by dependency

reads dependency

evaluates assurance

alternative to variability

Design guidance

Classification

Patterns
Capped phrase counting to prevent score inflation

Sources

  1. Ch3.1B: T. Nguyen, "Implement Evaluation Pipelines and Task Benchmarks - Guided Practice," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.1B. ISBN: 9798244538229.
  2. Ch4.2: T. Nguyen, "Deployment and Scaling," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 4.2. ISBN: 9798244538229.
  3. Ref7.07: E. Li, V. Bellotti, R. Kraus, and R. Kao, "Build a retrieval-augmented generation (RAG) agent with NVIDIA Nemotron," NVIDIA Technical Blog, Sep. 23, 2025. [Online]. Available: https://developer.nvidia.com/blog/build-a-rag-agent-with-nvidia-nemotron/