The claims follow-up rules live in one coordinator's head. She's on leave next month.

The situation

Where the operating knowledge actually lives

The clinic's claims process works — as long as the senior billing coordinator is in the building. She knows which payers deny CPT combinations that other payers accept, which denials are worth reworking within the 30-day window, and which write-offs need the practice manager's signature. None of it is in the EMR. Some of it is in a printed binder from 2022 that no longer matches how two of the three payers behave.

Scheduling boundaries are equally tribal. Which provider takes new Medicare patients, who blocks Fridays for procedures, how far out pediatric slots may be booked, and which referrals need records before an appointment is offered — front desk staff at each location answer these differently, and the differences show up as double-bookings, wrong-provider appointments, and patients told three different things.

The clinic wants to add an AI assistant for scheduling triage and claims-status drafting. It cannot: there is no source of truth to point the assistant at. Feeding it the 2022 binder would automate the wrong answers.

The work

What Taz does with this clinic

Taz inventories every place operational policy currently lives — binders, EMR macros, sticky notes, email threads, and structured interviews with the people who hold the rules — then converts it into a portable, structured bundle. The clinic reviews and approves every stated rule before it becomes canonical.

• Policy inventory: Every operational rule found is logged with its current location, owner, last-verified date, and confidence level. Conflicts between locations are surfaced as explicit decision items for the practice manager, not silently merged.
• Claims playbook extraction: Denial-handling rules are captured per payer: rework windows, documentation requirements, escalation thresholds, and write-off authority. Structured interviews with the billing coordinator turn her judgment into written, reviewable criteria.
• Scheduling boundary map: Per-provider rules — panel status, referral prerequisites, booking horizons, procedure blocks — become structured records the front desk at any location can query and any scheduling assistant can respect.
• Access boundary documentation: Who may view, book, or modify what — by role, not by name — so the rules survive staff turnover and can gate any future AI integration.
• OKF packaging: Everything becomes markdown concept files with YAML frontmatter, an index for progressive disclosure, cross-links between related policies, and a maintenance checklist naming who re-verifies each policy and how often.

The deliverable

What the clinic receives

• Policy inventory spreadsheet — every rule found, where it lived, who owns it, its verification status, and every inter-location conflict flagged for a decision.
• Claims operations bundle — per-payer denial playbooks, rework criteria, escalation thresholds, and write-off authority in structured markdown a biller or an AI drafting assistant can follow step by step.
• Scheduling and referral rulebook — per-provider boundaries and referral prerequisites as structured records, consistent across all three locations.
• Access boundary matrix — role-based view/book/modify permissions, ready to inform EMR configuration and any future AI deployment.
• Index and maintenance checklist — the entry-point file that makes the bundle navigable, plus a named-owner review cadence so the bundle stays true after handover.