AI & Enhanced

Pulse — Healthcare

A workforce operating system for the rugged Zebra TC52 — predicting patient deterioration six hours ahead and coordinating care teams at the speed of the floor.

Daily active time on TC52 grew from 4 min → 38 min per shift in pilot.
~22 minutes saved per shift on handoffs & alert triage.
31% drop in overhead pages on Ward 4B.
Zebra TC52
9:41

Pulse

CLINICAL WORKFLOW · NURSING

Regional Medical Center · v4.2.1

Project Overview

Most customers buy the TC52 handheld for one job — scan a wristband, scan a med, done — and then the device sits in a charger for the rest of the shift. Our customer-success data kept showing the same thing: high adoption for scanning, near-zero adoption for everything we shipped on top.

Pulse started as an internal initiative to change that. Instead of pitching more features, we embedded with a pilot hospital's night shift for 14 weeks on Ward 4B to figure out what the TC52 should actually do between scans — and shipped a workforce OS spanning 26 screens and three phases of AI.

Zebra TC52
9:41

Pulse

CLINICAL WORKFLOW · NURSING

Regional Medical Center · v4.2.1

01 · Splash
Brand intro with hospital tenant, quiet loading state

The first second of the app — Pulse mark, hospital name, and a quiet loading state while the device pulls the user's shift context.

Nurse taps Pulse on the TC52 home screen at clock-in. Loads in under a second so it feels instant, not like another slow hospital tool.

Sets the tenant (which hospital, which unit) before login, so we don't ask the nurse anything we already know.

Zebra TC52
9:41

Welcome back

Sign in to view your shift assignments

RN-48217
••••••••
OR

Help desk · ext. 4040

HIPAA compliant · SSO active
02 · Secure Login
Badge SSO + employee ID fallback, built for gloved hands

Tap-to-badge SSO with employee ID fallback. No long passwords, no email field — built for gloved hands and shared devices.

At shift start the nurse taps their badge to the back of the TC52. If the badge fails, they punch in a 6-digit ID.

Removes the 30–60 seconds nurses lose to password resets, and keeps HIPAA-grade auth without slowing the floor down.

My Role in the Project

I owned design end to end: floor research, sketches, Figma prototypes, the full UI, contributions to our internal Mobility DNA tokens, and handoff to the Android team. The PM ran customer comms, two engineers built it, and our clinical SME — a former ICU nurse on our staff — kept me honest about workflow.

A lot of the job was unglamorous: negotiating with hospital IT to side-load past their MDM, re-shooting reference photos because the first batch were all right-handed, and sitting through 11 p.m. staffing huddles so I'd design for the floor that existed, not the one I imagined.

Challenges Encountered

  • →The TC52 is a 5" screen wrapped in a thick rubber bumper — iPhone-sized tap targets don't survive a gloved thumb.
  • →Staff use it one-handed, almost always while holding something else: a chart, an IV bag, a phone for a doctor.
  • →Every tap is audited. The usual undo patterns weren't available to soften mistakes.
  • →Alert fatigue wasn't about volume — a low telemetry battery and a deteriorating patient looked identical on the device.
  • →Handoffs were wildly inconsistent. The incoming shift quietly redid a lot of triage as a result.
  • →Trust gap: the hospital had been burned by a previous charting rollout and Zebra was the company that made the device staff resented carrying.

User Research Approach & Insights

I spent two weeks shadowing eight nurses across three wards, plus two night shifts I'd rather not repeat. Six handoffs observed, four contextual interviews with charge nurses, and a one-week diary study with twelve staff. I also pulled six months of anonymized device telemetry from our own backend — which app was open, for how long, where the device was physically docked — and that turned out to be as useful as anything I heard in person.

Methods

  • Contextual inquiry: shadowed 8 nurses across 3 wards (incl. 2 night shifts).
  • Semi-structured interviews: 4 charge nurses, 6 floor RNs.
  • Diary study: 12 staff over one week.
  • Telemetry analysis: 6 months of anonymized device usage.
  • Stakeholder workshops with the clinical SME and Mobility DNA team.

Key Findings

  • →One charge nurse spent 12 minutes hunting a float RN who was "on shift" — gave up and paged overhead.
  • →A nurse's personal notebook was more trusted than the official handoff tool.
  • →Alert fatigue was a hierarchy problem, not a volume problem.
  • →Staffing was reactive — "we need help" came an hour late.
  • →Nurses didn't want AI. They wanted fewer steps. "Smart" had to earn it in two uses.

Ideation Process

1Discovery

Floor shadowing, charge-nurse interviews, telemetry audit.

2Ideation

Crazy 8s, journey maps, an effort/impact roadmap.

3Wireframing

Paper TC52 frames, one-thumb reach studies.

4Prototyping

Figma flows wired to a TC52 emulator on real hardware.

5Testing

Gloved-hand tap testing, on-floor pilot with 6 nurses.

6Iteration

Reordered actions, tuned contrast, killed three features.

The Solution Implemented

Design standards

  • 44px minimum tap targets, 56px for anything destructive or time-sensitive — above Mobility DNA defaults, because gloves.
  • Calm palette — soft lavender to sky. Wards are visually loud; the app should feel like a deep breath, not another alarm.
  • Every AI suggestion is editable, attributable, and easy to ignore.

Three phases layered on a core workforce app

  • Core: tasks + scanning
  • Phase 1: shift-summary / alert triage / staffing forecasts
  • Phase 2: prediction + voice agent + command center
Core

Core Experience

Tasks, scanning, and the foundation everything else sits on.

Zebra TC52
9:41
EC

Good morning

Emma Chen, RN

● On shift · Ward 4B

7

Code Blue · Room 412

Cardiac arrest · 2 min ago · 3 responders

Active Tasks

3 due

12

Staff on floor

Live

8/14

EM
JS
RK
AL
+4

Staffing suggestion

Charge nurse

Reassign 2 cardiology tasks to James Patel, RN — workload balanced and ACLS-certified.

Today's census

View all

7:00 AM – 7:00 PM

68% occupancy
7a10a1p4p7p

FRI

21

Next shift · ICU nights

7:00 PM – 7:00 AM · with Dr. Patel

01 · Dashboard Home
Alerts, workload, AI insights

The shift cockpit. Open alerts, today's workload, and a short AI read on what needs attention next — all above the fold.

Nurse opens this every time they unlock the device. It's the home base they come back to between rooms.

Replaces the mental juggling of pager + whiteboard + EHR tabs with one calm view of 'what matters right now'.

Zebra TC52
9:41

My tasks

12 active · 3 due today · 1 overdue

STAT· Rm 412B

Administer insulin · Walker, R.

due in 8 min
E
HIGH· Rm 308

Pre-op vitals · Martinez, J.

09:45
60%
E
HIGH· Rm 215A

Wound dressing change · Lee, S.

10:30
30%
E
MED· Rm 4B

Medication reconciliation · 4B

11:00
80%
E
LOW· Rm Hall

Restock supply cart 4

14:00
E
02 · My Tasks
Priority-coded, progress tracking

A personal task list color-coded by urgency, with a clear sense of what's done, what's in progress, and what's overdue.

Nurses check between patient rounds; charge nurses glance at it to see how a teammate's shift is tracking.

Ends the 'sticky note on the COW' workflow. Nothing falls through the cracks at handoff.

Zebra TC52
9:41

New task

Dictate, type, or a template

Priority

Assign to

SK

Sarah Kim, RN

Cardiology · Available · Ward 4B

Suggested
Workload 62%· Cert ACLS· Shift ends 19:00

Due

Today, 11:30

Patient

Walker, R. · 412B

Attachments

03 · Create Task
Voice, photo, smart assign

Add a task in seconds using voice dictation, a quick photo, or a tap. Pulse suggests the right assignee based on skills and location.

A nurse spots a spill or a missing supply, dictates it on the go, and Pulse routes it to the nearest qualified hand.

Cuts the friction of 'who do I call for this?' — the request gets to the right person without a phone tree.

Zebra TC52
9:41

Find staff

14 on shift · 8 available now

Search by name, certification, ward…ACLS

Availability · next 12 hours

07:00 – 19:00
7aLow
High7p
SK

Sarah Kim, RN

Cardiology · Ward 4B

62%
JP

James Patel, MD

Emergency · ED bay 3

48%
AL

Aisha Lopez, RN

ICU · Ward 6

88%
MW

Marcus Webb, CNA

Med-Surg · Ward 3A

35%
LT

Linh Tran, RN

Pediatrics · Ward 5

0%
04 · Associate Finder
Heatmap + skill search

Search staff by skill, role, or certification with a live heatmap of where they are on the unit.

Charge nurse needs an IV-certified RN now — types 'IV', sees who's available and closest.

Replaces the 12-minute walk-and-ask hunt for a specific clinician.

Zebra TC52
9:41

Floor 4B

Live staff and bed status

411A412A413A414A415A411B412B413B414B415BSJAML

Live · 5 staff

10 of 10 beds

Nearest available

SK

Sarah Kim, RN

Rm 408 · ACLS · 12 m walk

Available
Busy
Critical
Station
05 · Availability Map
Live floor view, nearest staff

A floor-plan view of the unit showing rooms, staff positions, and who's free vs. tied up.

When a code or a rapid response hits, the charge nurse opens the map and sees who can move first.

Gives the unit a shared sense of 'where is everyone right now' without anyone having to radio in.

Zebra TC52
9:41

Pulse Assist

Ask about staff, patients, shifts

Which cardiology staff are free in the next two hours?

Three cardiology staff are available between 10:30 and 12:30:

SK

Sarah Kim, RN

Available now · Ward 4B

ACLS
DC

David Cho, MD

Free in 25 min · Clinic

Echo
MS

Maya Singh, NP

Available · Cath Lab

PCI

Just now · from staff roster

Summarise active alerts in the ICU.

4 active alerts across 12 ICU beds:

  • ● Rm 412 · BP drop, rapid response active
  • ● Rm 408 · SpO₂ 91%, monitoring
  • ● Rm 403 · Pain score 7, PRN due
  • ● Rm 410 · Lab results posted
Ask a question…
06 · AI Assistant Chat
Conversational workflow

A chat thread with the Pulse AI — ask for a patient summary, request a task, or look up a policy in plain English.

Nurse types 'summarize bed 12 since last shift' while walking to the room.

Pulls answers out of the EHR without forcing the nurse to dig through tabs on a 5-inch screen.

Zebra TC52
9:41

Listening

"Assign medication review to Sarah"

Captured

Task: Medication review

Assigned to: Sarah Kim, RN · Cardiology

Priority: High · Due in 30 min

07 · Voice Assistant
Hands-free dictation

Push-to-talk voice input with a live transcript. Designed for moments when both hands are occupied.

Mid-procedure, the nurse holds says 'Hey Zebra, log 250 mL output for room 8'.

Lets clinicians document in the moment instead of stacking up charting for the end of shift.

Zebra TC52
9:41

Inbox

7 new alerts · 2 critical

Now

!

Code Blue · Room 412

2m

Cardiac arrest · 3 responders en route

℞

Medication due · Walker, R.

8m

Insulin lispro 10u SC · window closes in 8 min

Earlier today

⌚

Shift handoff in 30 min

30m

Report to night team · ICU floor

📢

Hospital-wide announcement

1h

Updated PPE protocol effective tomorrow

✓

Task completed by Sarah Kim, RN

2h

Pre-op vitals · Martinez, J.

♥

Lab results posted · Chen, M.

3h

Troponin within normal limits

08 · Notification Center
Priority-grouped feed

All alerts in one place, grouped by urgency rather than time — critical at the top, FYI at the bottom.

After a long room visit, the nurse opens this to triage what came in while they were away.

Fixes alert fatigue: instead of 40 dings, the nurse sees the three things that actually need a decision.

Zebra TC52
9:41

Announcements

From hospital leadership

PINNED
▶ 2:14

Quarterly town hall · Q3 results and 2026 plan

Dr. Eleanor Briggs, CMO · Recording available

LeadershipAll staff
POLICY· Today, 08:00

Updated hand hygiene audit schedule

Infection Prevention

EVENT· Yesterday

Nurses Week luncheon · Friday, 12:00

HR Engagement

DRILL· Mon

Code Silver drill · Building C, 15:00

Security Office

UNIT· Mon

New cardiac monitors deploying next week

Cardiology

09 · Announcements
Leadership + rich media

A clean feed for unit and hospital-wide announcements with images, video, and acknowledgements.

Manager posts a new isolation protocol; staff read and tap to acknowledge during a break.

Replaces the printed huddle sheets and forwarded emails nobody reads.

Zebra TC52
9:41

Patient wristband

METRO GENERAL

CHEN, MARCUS

DOB 04/12/1968 · MRN 408192

Align wristband within frame

Manual entry

10 · Patient Scanner
Wristband barcode verify

Uses the TC52's barcode scanner to read a patient wristband and pull up that patient instantly.

Before any med pass or procedure, the nurse pulls the trigger and scans the wristband to confirm identity.

Hard-stops wrong-patient errors — the single biggest source of preventable harm in med admin.

Zebra TC52
9:41
RW
!

Walker, Robert

M · 67 y · MRN 4827193 · DOB 14/03/1958

⚠ Penicillin allergyRm 412B

BP

128/82

HR

76

SpO₂

94%

Temp

99.1°

Heart rate · last 6 hours

6h ago3h1hNow

Active medications

Metoprolol 25 mg PO BID

Next: 12:00

Atorvastatin 40 mg PO QHS

Next: 21:00

Insulin lispro 10 u SC

Due in 8 min

11 · Patient Details
Vitals, meds, allergies

A patient summary built for a 5-inch screen: latest vitals, active meds, allergies, and recent notes — no horizontal scrolling.

Opens straight from a scan or alert. Nurse reads it on the way into the room.

Replaces the 'open the EHR on a workstation' detour for the 80% of lookups that are quick checks.

Zebra TC52
9:41

Assignment rationale

Why this nurse was suggested

EM

Suggested for

Emma Mason, RN

Task: Post-op cardiac assessment

5/5

Criteria met

Matching criteria

♥

ACLS and PALS certified

100%
⏱

On-shift, available within 15 min

92%
📊

Current workload 38% — lowest on unit

86%
✓

No active medication-error flags

96%
👤

Prior care for this patient (3 shifts)

74%
⚠

Handoff risk

Emma's shift ends in 2 h 40 min. Plan a handoff before assigning.

12 · Explainable AI
Decision factors + confidence

Shows why the AI recommended what it did — the top factors, the confidence level, and what data it used.

When Pulse flags a patient as deteriorating, the nurse taps 'why?' and sees the three vitals trends behind the call.

Builds the trust clinicians need to act on AI. Without this, nurses (rightly) ignore the model.

Zebra TC52
9:41

Analytics

Cardiology · this week

Task completion

94%

+3.2%

Avg response time

2.4m

−18%

Staff utilisation

UsedIdle
SK

Sarah Kim, RN

88%
JP

James Patel, MD

72%
AL

Aisha Lopez, RN

95%
MW

Marcus Webb, CNA

58%
LT

Linh Tran, RN

34%

Patient alerts by hour

3 critical
00:0006:0012:0018:0024:00

Staffing note

Aisha has been at 95% utilisation for 3 days. Consider redistributing her caseload.

13 · Analytics Dashboard
Utilization & alert trends

Personal and unit-level analytics — task throughput, alert response times, time-on-feet trends.

Nurses check their own numbers at end of shift; managers use it for weekly huddles.

Turns gut-feel about workload into something you can actually point at and fix.

Zebra TC52
9:41
EC

Emma Chen, RN

BSN · ACLS · PALS

CardiologyICU-certified5 yrs tenure

247

Tasks this wk

96%

On-time rate

4.9

Peer feedback

This week

Shifts completed

5 of 5
✓
M
✓
T
✓
W
✓
T
✓
F
–
S
–
S
🎓

Certifications and training

3 expire within 90 days

⏰

Shift preferences

Days · 12-hour

🔔

Notification settings

Critical only

🔒

Privacy and audit log

HIPAA access log

14 · User Profile
Stats, certs, preferences

The nurse's own page — certifications, shift history, notification preferences, and accessibility settings.

Updated when someone earns a new cert or wants to silence non-critical alerts on night shift.

Keeps the system's picture of each nurse accurate so smart-assign and alert routing actually work.

Phase 1 · Fastest ROI

Phase 1 · Fastest ROI

Shift-summary, handoff, alert ranking, and staffing forecasts.

Zebra TC52
9:41

Shift summary

AI Nurse Copilot · 7a – 7p

Copilot ready

12 patients · 18 events ingested

SBAR

Output · Marcus Chen, Rm 412

Status

Patient stable throughout shift.

Labs

Troponin pending (drawn 14:22).

Vitals

BP within range. HR 78 avg.

Pain

No chest pain reported since 09:00.

Plan

Cardiology recommends continued monitoring.

Generated in 1.2 s · GPT-4o · grounded on 18 chart events

15 · AI Shift Summary
Nurse Copilot · auto SBAR

An auto-generated recap of the nurse's shift in SBAR format — situation, background, assessment, recommendation.

Nurse opens this at the end of shift, edits anything that's off, and ships it to the next RN.

Saves 15–20 minutes of writing per shift and makes handoffs more consistent across the unit.

Zebra TC52
9:41

Shift handoff

Auto-built SBAR · ready to send

MC

Marcus Chen

M · 58 · Rm 412 · Telemetry

Watch
S

Situation

Telemetry monitoring post-ED arrival for atypical chest pain.

B

Background

58y M, HTN, prior NSTEMI 2023. Recent chest pain x2 episodes today.

A

Assessment

Stable vitals. ECG no acute ST changes. Pain free since 09:00.

R

Recommendation

Continue telemetry. Await troponin. Page cardio if >0.04.

16 · AI Handoff Generator
One-tap SBAR handoff

Takes the shift summary one step further: assembles a per-patient handoff packet ready to read out loud.

During shift change, the outgoing nurse walks the room with their TC52 and reads from this.

Fixes the inconsistent, memory-based handoffs that are the root cause of many sentinel events.

Zebra TC52
9:41

Alerts

AI ranked · 15 → 3 actionable

Noise reduced 80%

12 duplicates merged · 3 deferred

Auto

Critical now

· 1
STAT

Room 412 · Cardiac arrest

Code Blue · 30s ago

High priority

· 2
HIGH

Telemetry review overdue · Rm 402

Due 8 min ago

HIGH

Insulin recheck · Walker, R.

Sliding scale due

Can wait

· 2
INFO

Policy update · Infection control

For review by EOD

LOW

Cart 4 restock

Threshold reached

17 · Smart Alert Prioritization
15 alerts → 3 actionable

Re-ranks incoming alerts based on patient context, history, and current acuity — collapses noise into a short action list.

Replaces the raw alarm feed as the default notification view for floor nurses.

Alarm fatigue. Nurses stop ignoring the screen because what's on it is actually worth looking at.

Zebra TC52
9:41

Staffing forecast

Next 12 hours · 89% confidence

ER shortage likely · 5p – 8p

Predicted: 2 trauma-certified RNs short

Demand vs capacity

SHORTAGE
2p4p6p8p10p
Staffed
Predicted demand

Recommended actions

Call in 2 trauma RNs from float pool

-72% shortage risk

Extend Sarah Kim's shift by 4h

-31% shortage risk

Defer 3 elective admits to Saturday

-18% shortage risk

18 · AI Staffing Forecast
12-hour demand prediction

A 12-hour forward forecast of unit demand by hour, with a confidence band and the drivers behind it.

Charge nurse and house supervisor use it at huddles to decide on float pulls and break coverage.

Moves staffing from reactive ('we're drowning') to proactive ('let's pull one float at 2pm').

Phase 2 · Differentiator

Phase 2 · Differentiator

Prediction, autonomous tasks, and clinical voice agent.

Zebra TC52
9:41

Clinical risk

Marcus Chen · Rm 412 · refreshed 2m ago

72

/100

MEDIUM RISK

Deterioration likely within 4 h

+18 pts since 06:00

Top contributors

↑

HR trend rising

82 → 104 over 3h

↓

SpO₂ variability

σ 2.1 → 4.8

!

Recent cardiac alert

PVC run · 14:02

Suggested actions

Repeat vitals in 30 min

Notify charge nurse

Confirm cardiology consult ETA

19 · Predictive Deterioration
Clinical risk score + drivers

An early-warning view that scores each patient's risk of deteriorating in the next 6 hours, with the trends driving the score.

Charge nurse scans this every couple of hours; rapid response teams use it on their rounds.

Catches the silent declines that today only show up once vitals crash.

Zebra TC52
9:41

Auto-tasks

AI creates · you approve

New order detected

Troponin · Chen, M · 14:22

Auto
AI· in 10 min

Blood draw · Rm 412

Assign nearest phleb

AI· in ~45 min

Lab follow-up

Watch result · auto-page if >0.04

AI· in on result

Notify Dr. Patel if elevated

Pager + SMS

AI· in on result

Document in chart

Auto-template SBAR

Saves ~7 min per order · 142 tasks auto-created this week

20 · Autonomous Task Creation
AI proposes · nurse approves

When the AI sees a pattern that needs action — a missed turn, a due med — it drafts a task and waits for a one-tap approval.

Nurse reviews proposed tasks at the start of shift and at natural pauses through the day.

Closes the gap between 'the EHR knows' and 'someone actually did the thing'.

Zebra TC52
9:41

Clinical voice

HIPAA · on-device

You

"Hey Zebra, assign telemetry review to nearest cardiology-certified nurse."

Pulse · executed

Assigned Sarah Kim, RN · ACLS · 12 m from Rm 402. Pager sent.

Skill ✓Distance ✓Load ✓

You

"Hey Zebra, summarize Room 402."

Pulse · speaking

"Stable. Troponin pending. No chest pain. Telemetry active."

0:08

Tap to interrupt · "Hey Zebra" to wake

21 · Clinical Voice Agent
Speak commands · AI executes

A conversational agent that can run multi-step actions — log vitals, place a request, summarize a chart — entirely by voice.

Nurse, hands full at the bedside, says 'Hey Zebra, log BP 142 over 88 for this patient and flag the on-call'.

Removes screen-time from the bedside. The nurse stays with the patient, not the device.

Phase 3 · Enterprise

Phase 3 · Enterprise

Command center, executive briefings, and resource optimization.

Zebra TC52
9:41

Shift command

Charge view · live ops

AI OPERATIONS

ICU → 97% workload by 6 PM

Threshold breach in 3h 12m

Recommended actions

Move 1 float RN (Ward 4B → ICU)

Reassign Rm 402 telemetry review

Delay non-urgent wound checks

Expected improvement

-18% response time

Unit pulse

ICU

97%

ER

84%

4B

62%

PCU

71%

22 · AI Shift Command Center
Live ops alerts · charge view

A house-wide live view of every unit's load, staffing gaps, and active escalations — built for charge and house supervisors.

House supervisor keeps this open on a tablet at the desk and on their TC52 when walking.

Replaces the whiteboard and the four phones it takes to keep a 200-bed hospital coordinated.

Zebra TC52
9:41

Hospital pulse

Week 25 · executive briefing

For

Dr. Helen Voss · CNO

AI · 6:00 AM

Strong week overall. Floor responsiveness improved, but ICU staffing pressure builds toward Friday.

Response times

-14%

vs last wk

Task completion

+9%

4,218 closed

ICU staffing

92%

threshold near

Patient sat.

4.7

+0.2

Forecasts

ER overload predicted

Friday 5p – 9p

ICU near staffing threshold

Thu – Sat

Elective surgery surge

Next Mon

Top recommendation

Increase float pool by 2 staff next week. Projected ROI: -$48K overtime.

23 · Executive AI Briefing
Weekly pulse for leadership

A weekly auto-generated briefing for the CNO and C-suite: throughput, safety events, staffing health, and what changed.

Monday morning read for nursing leadership before their ops meeting.

Turns a dashboard nobody opens into a short narrative leaders actually read and act on.

Zebra TC52
9:41

Resource optimization

Hospital-wide · live

$142K saved this month

+11% bed utilization · -23% agency spend

Live rebalancing

4B
ICU

1 RN

ICU census spike

OR-3
PACU

2 beds

Surgery delayed 2h

Float
ER

1 trauma RN

Predicted surge 5p

Optimization targets

Bed utilization

87 / 92

OR turnover (min)

24 / 18

Avg LOS (days)

4.2 / 3.8

Float-pool yield

78 / 85

24 · Resource Optimization
House-wide rebalancing

Suggests moves — float nurses, equipment, beds — to rebalance the house when one unit spikes.

House supervisor reviews suggestions at shift change and accepts the ones that make sense.

Stops one unit drowning while another sits quiet, without endless cross-floor phone calls.

Results & Impact

Six-week pilot on Ward 4B.

Daily active time on TC52 outside scanning: 4 min → 38 min (device telemetry, n=14).

~22 minutes saved per shift on handoffs and alert triage (self-reported, n=14).

31% drop in overhead pages on Ward 4B vs. the prior 4-week baseline (paging-system log).

Charge nurse retired her paper staffing grid by week 3.

Rolling out to 4 more wards this quarter; 2 new health-system customers in early talks.

Three nurses, unprompted: "this feels like it was built by someone who's been on the floor."

Key Learnings & Reflection

Key Learnings

Working at the company that makes the hardware is a cheat code and a trap — I had to resist designing for the spec sheet instead of the human holding it.
Accessibility isn't a checklist at the end. On a TC52 at 3 a.m. with one hand busy, the whole product is an accessibility problem.
Cutting features was harder than designing them. We killed a social feed, a leaderboard, and a dark mode I personally loved — none rooted in anything a nurse asked for.
Spend time with the people you're designing for, then spend more. Every good decision traces back to something I saw on a shift.
Trust beats novelty in clinical software — every model-generated suggestion needs a one-tap override or nurses (rightly) route around it.

Reflection

"Designing Pulse for the TC52 reframed what 'mobile' means in healthcare for me. The win wasn't a smarter algorithm or a prettier screen — it was earning a place in the nurse's actual workflow, between rooms, between glances, between gloves. When the staff who'd called the old tool 'the Tax' started reaching for Pulse first, the project landed."