Healthcare & Clinics

The waiting room emptied before the doctor arrived

Four clinics moved patient intake to the phone in the waiting room and got eleven hours a week back at reception.

This is an illustrative composite. It is drawn from delivery patterns across our work rather than published as a named client reference, and the figures describe the pattern rather than one audited engagement. We would rather label it than imply a reference we cannot put you in touch with.

Size
85 staff
Platforms
ConnectAtlas

Context

An 85-staff group running four clinics with a shared reception pool. Intake was a clipboard, re-keyed into the practice system by whoever was on the desk, usually twice — once at registration and again when the doctor needed the history.

The problem

Reception spent most of the morning typing. Re-keying introduced errors in exactly the fields that matter — allergies, current medication, insurance number — and the correction loop ran through the doctor. Peak queue at the 9am opening was regularly twenty minutes before anyone was seen.

What we did

  1. 01Sat at two reception desks through a full morning at each of the four clinics. The workarounds at the busiest clinic were the requirements.
  2. 02Moved intake to a WhatsApp thread on the number the patient already books from, with the form adapting to the appointment type.
  3. 03Validated the high-risk fields against the practice system on submission rather than on arrival, so a mismatch surfaces the day before.
  4. 04Gave the doctor a structured summary at the top of the record, with every line citing the answer it came from.
  5. 05Left the clipboard in place for walk-ins, because removing it would have made the exception the hardest path.

Architecture

Nothing here is a black box. You can read the architecture before you sign, and you own it after.

  1. 01ChannelJJ Tech Connect on WhatsApp Business, threading each intake to the booking it belongs to.
  2. 02RecordIntake responses in Postgres, org_id per clinic, deny-by-default policies with per-role read.
  3. 03ValidationZod schema at the boundary; insurance and medication fields checked against the practice system before arrival.
  4. 04SummaryJJ Atlas produces the clinical summary with a mandatory citation per line — a claim with no source does not render.
  5. 05OversightThe summary is a draft until a clinician accepts it. No generated text enters a record unreviewed.

Results

The numbers the work closed against.

target reception time returned across four clinics
11 hrs/weektarget reception time returned across four clinics

Basis: Arithmetic: observed re-keying time per intake multiplied by weekly intake volume across four sites, assuming the 78% pre-arrival completion target below. Walk-ins stay on the clipboard and are excluded.

target median wait at the 9am opening
20 min → 6 mintarget median wait at the 9am opening

Basis: 20 minutes is observed. The target assumes registration collapses to identity confirmation once intake arrives complete, leaving only the queue the desk cannot compress.

target share of intakes completed before arrival
78%target share of intakes completed before arrival

Basis: Assumes patients who book through the messaging channel complete there. Derived from the share of bookings already arriving on that channel, discounted for patients who do not respond.

generated summary lines entering a record without clinician acceptance
0generated summary lines entering a record without clinician acceptance

Basis: A structural constraint, not a forecast: the summary is a draft object and has no write path into the record without a recorded acceptance.

Stack

  • Next.js
  • Supabase Postgres
  • WhatsApp Business API
  • pgvector
  • Zod
  • Vercel

Related work

60 staff

Reports that stopped needing a chase

A diagnostics lab cut report turnaround from three days to four hours by fixing the handoffs, not the analysis.

4 hours target report turnaround, against a three-day baseline

28 staff

Weekend enquiries answered before Monday

A lettings agency stopped losing the enquiries that arrived when nobody was at a desk.

8 min target median first reply to a weekend enquiry

Does this look like your business?

Tell us what is slow and what it is costing. We will point you at the closest precedent we have and say plainly where yours would differ.

Scope it with us