The Phone, the Portal, and the Empty Slot: A Patient Access Desk for a Family Practice

A family practice has two problems that look opposite and are the same problem. Patients cannot get through on the phone, and appointments go unfilled. At 8:55am the line has eleven callers waiting; by 3pm two of that afternoon’s slots are empty because somebody cancelled and nobody had time to work the list. The practice is not short of demand or short of capacity. It is short of the thing that connects them.

This is the transformation we would run for that practice. It is presented as a concept, built on the same method we use in client engagements, so the before, the after, and the path between them are all visible.

What does the access problem actually cost?

Practice leaders already know where this sits. In an MGMA Stat poll of medical practice leaders taken on 9 December 2025, with 236 applicable responses, no-shows were the top patient-access priority for 2026 at 27%, followed by online scheduling at 24%, phone access at 22% and wait times at 21%. Those four are usually described as separate projects. They are four symptoms of one thing: requests arrive faster than a person can classify them, so the queue and the calendar drift apart.

The self-service answer has already been tried and has largely not worked. An earlier MGMA Stat poll in July 2025 found that 71% of medical groups had fewer than one in four patients using digital tools to schedule appointments. The portals exist. Patients are not using them, because a booking portal asks the patient to do the practice’s classification work: choose a provider, choose a visit type, choose a duration. A patient with a lump on their neck does not know whether that is a fifteen-minute or a thirty-minute appointment, so they phone instead, and the queue grows.

Before and after: the front desk queue

Drag the handle. The before is the recreated day sheet and phone queue; the after is the triaged access desk designed in its place.

The phone does not disappear, and neither does the receptionist. What changes is what reaches them: a queue that has already been read, sorted and matched against real openings, with the reasoning shown so a person can overrule it in one click.

What we built

The concept above is the output of the same engagement steps we run in client work, and two of them are decisions a competent peer could argue with.

• We map the four intake channels first, phone, portal, fax and walk-in, and deliberately unify them at the queue rather than at the front end. Replacing the phone system is the obvious move and the wrong one: it is a long, expensive project with a change-management cost, and the practice would still be classifying by hand at the end of it.
• We design the request as a sentence, not a form. The patient writes or says what is wrong in their own words, and the classification happens behind the glass. This is the decision most likely to be challenged, because it means accepting unstructured input into a scheduling system on purpose.
• We leave urgency triage un-automated, and this is the deliberate gap. The system may propose a category and a duration, and it may rank the waitlist, but it never decides that a symptom is routine. Anything with a red-flag pattern, and anything the model scores below its confidence floor, goes to a human queue rather than to a category.
• We build the cancellation fill as a hold, never a booking. When a slot opens, the system ranks who on the waitlist fits it clinically and logistically, drafts the message, and holds the slot. A person releases it. Nothing is ever moved into a patient’s calendar without a staff member confirming.
• We wire evals from day one against the practice’s own historical requests, scored on whether the proposed category and duration matched what the clinician actually needed. Every staff override becomes a test case.

How it works

The EMR stays the source of truth, which is what makes this safe in a practice with real rosters and real billing. Requests from all four channels land in a single queue. Each one is read, given a proposed category and duration with the reasoning attached, and matched against genuine openings in the schedule. A staff member confirms or overrides, and only the confirmed booking writes back to the EMR. When a cancellation lands, the same queue ranks the waitlist against that specific slot and drafts the outreach, holding the appointment until a person releases it.

Because requests and chart context stay in Canadian infrastructure and processing runs with zero data retention, the design lines up with what PHIPA expects: a completed privacy impact assessment before deployment, an agreement naming the processor as an agent under section 17, and an audit log on every access to personal health information.

Why this pays back

The return arrives in capacity rather than in saved minutes, which is the number a practice administrator can take to the partners. Canada Health Infoway’s AI Scribe Program, whose six-month results were published in June 2026 across a national cohort of more than 12,000 enrolled primary care clinicians, found clinicians saved more than an hour a week during work hours, and that 20% of them redirected that time to increase appointment availability. That is the conversion this design is aimed at: not hours returned, but hours turned back into appointments.

Stated honestly, break-even for the first slice is about six recovered billable appointments a week at a typical Ontario family-practice visit value, and the practice should substitute its own number rather than trust that one. Here is the limit worth saying out loud before anyone signs: if the practice’s schedule is genuinely full, with no cancellations and no waitlist, this recovers almost nothing. It converts wasted capacity into used capacity, and a practice with no wasted capacity should spend the money on documentation instead.