Orchestration · Software component

Intent Router

Software componentOrchestrationOrchestration & Toolsarc:IntentRouter

A classifier agent that determines request intent and routing category, attaching a confidence score used for escalation.

Responsibility. Classifies requests into intent and routing categories.

Also known as: Classification Agent, Intent classifier, Query classifier, Categorisation node, Root decision node, Query domain recognition, Intent classifier service, Intent recognition, Intent detection

Direct neighbourhood (hover for relationship types)

Relationships

deployed on structural

Container Orchestrator Ch4.2

hosts structural

Intent Classifier Model Ch10.1

is configured by structural

invokes dependency

routes to dynamic

sends data to dynamic

Answer Synthesizer abstract Ch1.3
Escalation Agent Ch1.3

fails over to control

Rule-Based Intent Classifier Ch1.3

is orchestrated by control

is evaluated by assurance

Human Evaluator Ch10.1

is monitored by assurance

Intent Performance Analyzer Ch10.1

Design guidance

SHOULD use deterministic (low-temperature) inference for routing consistency.
SHOULD flag low-confidence classifications for escalation.
SHOULD constrain the classifier to return only the category name, reducing tokens, latency and parsing effort.
SHOULD use a low sampling temperature so routing is consistent rather than creative.
SHOULD enforce the set of valid categories at the type level in the state schema.
MAY be taught new service lines through few-shot demonstrations of correct routing decisions instead of retraining.
SHOULD be deployed and scaled independently from GPU-bound generation services.
SHOULD infer intent from conversation history, user behaviour patterns and domain knowledge, not surface keyword matching.
SHOULD map each intent to a distinct handling pathway (external API, RAG, transaction system, human escalation).

Quantitative guidance

As stated by the sources; verify before use.

Traditional LLM endpoints add 100-200 ms latency for simple classification; NIM-optimised serving reduced it to 30-50 ms (Ch1.5B).
Benchmark: classification latency ~150 ms reduced to ~50 ms, a 3x speedup (Ch1.5B).

Classification

Patterns: Confidence thresholdingLogic tree root decisionStructured-output classificationGoal inference beyond keyword matching
Technologies: LangGraphNVIDIA NIM
Quality attributes: Reliability (ISO/IEC 25010 | NIST AI RMF: valid and reliable)Functional suitability: correctness and validity (ISO/IEC 25010 | NIST AI RMF: valid)Performance efficiency (ISO/IEC 25010)
Risks mitigated: Misrouting of inquiriesMismatched responses from misrouted intents

Sources

Ch1.3: T. Nguyen, "Multi-Agent Systems," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.3. ISBN: 9798244538229.
Ch1.5B: T. Nguyen, "Stateful Orchestration - Worked Examples," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 1.5B. ISBN: 9798244538229.
Ch3.3: T. Nguyen, "Web Navigation and Interaction Benchmarks," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.3. ISBN: 9798244538229.
Ch3.5: T. Nguyen, "Prompt Optimization, Few-Shot Learning, Fine-Tuning," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 3.5. ISBN: 9798244538229.
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.
Ch10.1: T. Nguyen, "Conversational UI," in Mastering Agentic AI Systems: Guide for the NVIDIA NCP-AAI Exam, 1st ed. 2026, ch. 10.1. ISBN: 9798244538229.