Today
Team leads manually sample care team interactions from Zendesk and Assembled, score them against a rubric in a spreadsheet, and compile results for managers. Sampling is inconsistent and lags interactions by 1–2 weeks.
How the workflow helps
Tray can sample interactions from Zendesk and Assembled, score them against a defined rubric using LLM, surface results to team leads, and feed an ops quality dashboard — increasing coverage without adding reviewer headcount.
Value inputs to capture
- How many care team interactions per month are eligible for QA?
- How many minutes per item does manual scoring take?
- What sample rate does leadership want to reach?
Add volume, time saved, and loaded cost to create a draft estimate. Until then, no impact number is shown.