Experiment Center
Disciplined product iteration across the guided support workflow — illustrative results, not real performance data.
Guided support intake versus open chat
Workflow v2.4 · Started Aug 10, 2026
Structured intake improves handoff completeness without harming member satisfaction.
- Target member segment
- New members, first support conversation, all partners
- Success metric
- Handoff completeness
- Guardrail metric
- Member satisfaction
Handoff completeness is meaningfully higher with guided intake; satisfaction guardrail is holding steady, not yet statistically confirmed.
Approved-resource placement in AI summary
Workflow v2.4 · Started Aug 15, 2026
Surfacing approved resources above the recommended next step increases resource engagement.
- Target member segment
- All members with a completed guided conversation
- Success metric
- Resource click-through rate
- Guardrail metric
- Time to complete conversation
Directionally positive; sample size not yet sufficient for a confident read.
Plain-language rewrite of routing explanation
Workflow v2.3 · Started Jun 1, 2026
A simpler explanation of why a case was routed to a specialist reduces repeat contacts on the same topic.
- Target member segment
- Members whose request was routed to a specialist
- Success metric
- Repeat-contact rate
- Guardrail metric
- Member satisfaction
Plain-language rewrite reduced repeat contacts with no change to satisfaction. Rolled into Workflow v2.4.