AI agent quality, compliance, and access control evaluation dashboards
Full agent pipeline testing: tool usage, document retrieval, response correctness, and grounding across KNOWLEDGE, GREETING, and DECLINE scenarios.
Verifies company and role-based document access controls — GSC/APMT isolation, manager-only content, and cross-country filtering.
Standalone knowledge base retrieval quality. Tests document relevance, content matching, and response sufficiency using LLM-as-Judge scoring.
Tests that RAG retrieval respects country and company boundaries — India vs Global, GSC vs APMT, and cross-location isolation.