Finding What the Referral Left Out Before the Patient Arrives

What it does

Before a new patient intake appointment, the Care Coordinator needs to know what the referral contains and what is still missing. The material arrives through clinician messages, the scheduling record, prior visit notes, and structured fields such as eligibility codes. A gap found in the room rather than beforehand costs the patient a second visit.

The AI pulls the referral thread, the scheduling entry, and the accessible clinical summary for that patient, then compares them against the intake checklist the service operates. It produces a short intake brief, a list of open questions, and a source reference behind every retained fact. It flags an unresolved authorization, two different medication lists, and a stated urgency conflicting with the booked appointment type.

The Care Coordinator accepts, corrects, or discards each element of the brief. The Clinical Lead is involved only when a threshold applies: a judgment about clinical urgency or a request outside the scope of the service. Clinical and scheduling authority remains with the accountable human, and nothing enters the patient record without that action. Overrides are logged, so a checklist item rejected constantly can be revised by its owners.

Where the value comes from

Time spent chasing referral detail before an intake appointment is time the Care Coordinator cannot give the patients already waiting. Preparation minutes returned per intake is the benefit that carries the case, measured in the `handling_time_reduction` family. For a provider that shows first in clinics where incomplete referrals turn a booked slot into a rebooking.

To build the business case, the client measures three things: the annual intake volume in scope, the baseline preparation minutes per intake, and the fully loaded hourly cost of the coordinator. Those inputs drive the `transaction_based_time_saving` model, tempered by scope, adoption, and confidence factors. Acceptance and override data then show how much of the projected time came back.

The other benefit streams stay out of the headline. `search_time_reduction` describes lookup effort already counted, so including it would value one minute twice. `error_rework_reduction` reaches money through `error_rework_cost_avoidance`, but only after the provider fixes its own baseline for incomplete intakes and the cost of a repeat appointment. `knowledge_accessibility`, `user_experience`, `process_consistency`, and `exception_visibility` have no monetization path and no client inputs, so they remain control and adoption indicators. That separation keeps the claim defensible when finance reviews it.

How work changes

Routine lookup declines for the Care Coordinator. The role spends less time gathering messages, checking calendar context, rebuilding notes, and searching reference material. The assistant also reduces repeated reconstruction of context between meetings or shifts. Manual preparation around the case becomes shorter, especially after patient intake preparation.

Structured review increases for the Care Coordinator. The user must verify summaries, approve draft notes, maintain preference settings, and correct weak task suggestions. This shifts effort from collecting context to judging whether the prepared context is fit for use. The role also learns to mark weak outputs so future support improves.

Clinical Lead participates only when the assistant flags an item outside routine authority. That may involve policy exposure, safety concern, customer commitment, legal sensitivity, or commercial impact. The secondary role reviews the recommendation and confirms whether escalation is needed. That participation remains exception based, so daily ownership stays with the primary role.

Required skills include source checking, prompt discipline, exception recognition, and accountability for accepted outputs. Freed hours are monetizable only when managers redirect them to measured backlog, higher throughput, or avoided hiring. If time simply becomes slack, the business case should treat it as experience improvement. Team leads need clear capacity plans before counting the hours as financial value.