Unverified50% confidenceFactExact time
The researchers calibrated the Agentic Judge using 100 trajectories that were manually annotated call by call.
1
Sources
50%
Confidence
Long-term
Relevance
9/28/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe study is based on 100 lead qualification tasks in CRMArena-Pro.74% similarUnverifiedIn actual testing, one Deep Research call invoked 104 agents for targeted research, completing fact-checking based on 22 sources and 59 claims.72% similarUnverifiedThe analyzers executed by the policy engine hit the answer + evidence contract in all 110 out of 110 tests.71% similarUnverifiedThe research team conducted a technical audit of 200 samples to analyze the sources of discriminative signals.70% similarUnverifiedThe researchers ran the MARCH framework as-is on two code evaluation benchmarks, covering 80 condition-by-unit measurements.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/955463API
curl https://kongchang.com/api/v1/knowledge/claims/955463MCP
get_claim(id=955463)