Evidence · first-party tested

When the document lookup failed, the agent fell back to general knowledge without warning, so the failure was not surfaced transparently.

⚠ StruggledPickaxe →
Output — unretouched
tool output
The observation

When the document lookup failed, the agent fell back to general knowledge without warning, so the failure was not surfaced transparently.

Criterion: Observability · Scenario: Leave policy Q&A plus leave request draft for Arjun Desai (group: leave-policy)
Query this
get_evidence({
  tool: "pickaxe",
  scenario: "leave-policy"
})
MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Same input, other tools
Dust ✓ WorkedGumloop ✓ WorkedRelevance AI ⚠ StruggledZapier AI Agents ⚠ Struggled
Real outputs, no retouching · every cell queryable via API & MCP · aidemos.com