Session Objective

Build a repeatable QA audit for recurring treatment-status calls using RingCentral/RingSense transcripts, Litify Matter context, provider requests, notes/tasks, and ExactaCare/Aptiva Drive evidence where available.

What Was Accomplished

• Created scripts/build_treatment_status_call_qa_audit.py in the Litify analytics project.
• Created canonical skill Skills/ringcentral-treatment-call-qa/SKILL.md with slash command /ringcentral-treatment-call-qa.
• Registered the skill in Skills/SKILL-INDEX.md and Skills/LITIFY-SKILLS.md.
• Ran the live read-only audit against LITIFY_ORG.
• Wrote audit output to Repos/sail-litify/Litify_AI_Integration_Project/analytics/treatment_status_call_qa_2026-06-10/.
• Updated OPS_LEDGER.md and docs/litify-environment-diary.md with the live evidence and boundaries.

Live Evidence

• Source call rows: 1,549 Treatment-stage Matter-linked SAIL_Call_Log__c rows in the 60-day lookback.
• Sample: 40 matters, 5 per case manager, across 8 case managers.
• Workup evidence loaded: 1,226 Litify Notes, 194 Tasks, and 348 provider Request rows.
• Exacta/Aptiva cache reference: 27 provider rows from the June 4 connector-cache output.

Main Findings

• Best routine method: 5 active Treatment-stage clients per case manager, then review all calls in the lookback for those sampled matters.
• Recurring flags in the pilot: employment/missed-work not found in 21 cards, length/short-call flags in 8, progress flags in 7, provider verification flags in 6, and pain/what-is-hurting flags in 5.
• 33 cards had ExactaCare/Aptiva provider signals needing Drive/portal prework review, but this is a review prompt, not proof of staff failure, because the cache is not global.

Important Boundaries

• No Litify writes.