All concepts

Northline

Studio concept · UXProPatient booking · iOS

An iOS app for the patient side of a small Ontario clinic: book, prepare, attend, settle up. Privacy shaped the interaction, not just the fine print.

The 30-second version

Role
Design lead: framing, decisions, system, art direction
Team
Me, with AI agents building the 14 screens as HTML and generating the plates
Timeline
September 2026
Scope
14 screens in light and dark, a token set, a system sheet, 6 boards
What I decided
Consent is a screen in the flow. Today’s openings sit above the calendar. A six-day strip replaces the month grid.
Outcome
None claimed. A day-one measurement plan instead.

Read this first · ConceptStudio concept, self-initiated at UXPro. No clinic commissioned it, nothing shipped, nothing was measured. The clinic, practitioners, patients, benefits provider and every appointment and receipt are invented.

FIG. 000Home: the next visit, and the one thing left to do before it.

01 · The problem

Booking happens in business hours. The patient is at work.

Picture a clinic with four practitioners (physio, chiro, massage, pelvic health), one front desk and one phone line. Booking only happens in business hours, which is exactly when the patient is at work.

Intake is a clipboard five minutes before the appointment, so the practitioner reads it with the patient already on the table. Receipts go out as email PDFs, the channel Ontario’s PHIPA guidance says shouldn’t carry health detail without express consent.

The clinic’s real money problem is gaps. A no-show at 4:15 on a Tuesday can’t be recovered, and the only fix is a reminder the patient actually sees.

FIG. 001Booking at 10 p.m., when the front desk is closed.

02 · The person

Noor, 34, and a back she’s tired of explaining.

Desk job in Kitchener, recurring low-back pain, benefits with a physio cap she can’t remember. She wants to book at 10 p.m., know what she’ll pay before she walks in, and not re-explain her back to the third practitioner.

Behind her: a front desk whose day improves when intake arrives early, and a practitioner who’d rather read two good sentences than a page of boxes.

  • Constraints: PHIPA consent rules, extended-health direct billing, iOS conventions (44 pt targets, Dynamic Type to XXL), and one accent colour that had to read calm, not clinical-bright.

03 · Decisions

Six decisions, one moment of delight.

Every decision here is about reducing a specific worry: what will this cost, who will see this, will I remember.

  1. 01

    Consent is a screen in the flow, not a checkbox under a wall of text.

    Three toggles in plain language: care inside the circle of care (required), text reminders (on), clinic news by email (off). The copy names the real risk (“Ordinary email isn’t encrypted, so we keep health details out of it entirely”) instead of citing a statute at a patient. That screen set the voice for every privacy line in the app.

  2. 02

    Same-day openings sit above the calendar.

    Booking still runs treatment → practitioner → time, but a strip of today’s real openings leads, because the person with the highest intent is the one in pain today.

  3. 03

    A six-day strip replaced the month grid.

    Each day carries a load dot. A slot taken while you’re looking stays in place, struck through, so the grid never reflows under a moving thumb.

    Trade-offYou lose the at-a-glance month. On a phone, for a physio appointment, that’s a month nobody needed.

  4. FIG. 002Left: the strip in hand; the 2:30 slot was taken while you looked, so it stays put, struck through. Right: the confirmation rings, the strip up close, and the Live Activity.
  5. 04

    Intake is progressive, and openly abandonable.

    Section 3 of 5, “~2 min left”, a Saved pill, and “Finish later — your answers are saved”. Telling people they can stop is what makes them continue. The free-text prompt is personal: “Anything Daniel should know before Tuesday?”

  6. FIG. 003Book, prepare, arrive: the core flow in five steps.
  7. 05

    The day-of reminder is a Live Activity.

    Countdown, leave-by time, drive time and directions on the lock screen, with an ordinary notification for address and parking. It’s the one place a no-show gets prevented.

  8. 06

    Direct billing is a first-class flow.

    The insurer-paid and patient-paid halves of a receipt are shown separately, because that’s how an Ontario claim actually settles.

One moment of delight, and only one: five rings ease out from the confirmation checkmark and settle. Every other piece of motion reports state.

04 · The system

One family, one teal, dark mode as a token swap.

Instrument Sans, one family, tabular numerals throughout, because the app is mostly times, doses and dollars. A ten-step scale aligned to iOS (17 pt body, 34 pt large title), 8-pt spacing, 44 pt targets.

A single deep teal (#0C5C55) on warm off-white (#F7F4EF), quieter than the category’s bright teal. Every text pair was checked by script: tertiary ink went from #727D7B to #636E6C to take captions from 3.88:1 to 4.81:1. The teal next-visit card keeps its own dark gradient in both themes; it’s the clinic’s voice and shouldn’t invert.

FIG. 004The whole surface: fourteen screens, consent through to the error state.
  • System poster
  • Light & dark
FIG. 005The system on one sheet, and the same home screen after the token swap.

05 · If it shipped

What I’d measure if a clinic shipped it.

No numbers exist yet. These are the ones that would tell me which decisions were wrong.

  • Share of bookings started from the same-day strip.
  • Step-by-step booking drop-off.
  • Intakes finished before the visit (today’s baseline: at the desk).
  • No-shows with a Live Activity vs. SMS alone.
  • How many people keep text reminders on, and how many change consent later: a proxy for whether they understood it the first time.

06 · Reflection

Once privacy was plain sentences, the voice followed.

The consent screen was the decision everything else hung off. Next I’d put the consent and intake screens in front of five real patients, since those are the two places a concept is most likely to be wrong.