Quality Evaluation Score Discrepancy in EU-West BYOC WFM

Why does the calculated quality score in the Performance Dashboard diverges from the raw evaluation data exported via the WFM API?

We are operating in an EU-West BYOC environment. The business requires strict alignment between the agent performance view in the dashboard and the underlying evaluation records. Currently, agents are seeing a 94% score on their performance tile, while the exported JSON payload indicates a weighted average of 89%. This discrepancy is causing significant confusion during coaching sessions.

The flow architecture is standard, with no custom data actions modifying the score post-evaluation. The issue appears specific to the ‘Compliance’ category weightings.

Here is the payload returned for a single evaluation instance:

{
 "evaluationId": "eval-998877",
 "agentId": "agent-123",
 "categories": [
 {
 "name": "Compliance",
 "weight": 0.4,
 "score": 85
 },
 {
 "name": "Soft Skills",
 "weight": 0.6,
 "score": 92
 }
 ],
 "totalScore": 89.2
}

The dashboard displays 94. Is there a known caching delay or a specific configuration in the Quality Management settings that overrides the weighted calculation for display purposes? We need to ensure the metrics reported to leadership are accurate.