From Phone Queues and PDFs to One Calm Front Door for Care
A patient engagement app for a provider network: scheduling, visits, prescriptions, lab and pharmacy journeys, documents, and notifications—so people spend less time navigating the system and more time getting care.
The problem: Meet Daniel Okonkwo, VP of Patient Access at a multi-site care network. His clinicians were respected; his contact centre was underwater. Patients did not fail to care—they failed to find the next step. A lab order lived in one message thread, a pharmacy coupon in another, and the radiology prep sheet in a fourth. The product question was not “can we ship an app?” It was: can a family complete a care path without feeling like they joined an escape room?
The access metrics nobody brags about:
“We didn’t need another brochure with a logo,” Daniel says. “We needed a single front door that still respected how clinicians and fulfilment partners work.” The breaking moment was peak flu season: great medical outcomes, brutal wait times, and reviews that blamed ‘the process,’ not the doctors.
The mandate was a patient-grade mobile product on top of existing services: authenticated access, real schedules, and traceable actions—not a wrapper around a desktop-only portal.
What we built
Get care on the calendar—without the phone maze
- Search doctors, locations, and slots with clear confirmation and reschedule paths
- Reminders patients actually opted into—not generic blast spam
- Video visit entry where the programme offers it, beside in-person care
Continuity after the visit: prescriptions, labs, pharmacy
- Prescription history and next steps visible from the home journey
- Lab catalogues and results handoffs without “check your email for a PDF”
- Pharmacy exploration tied to the same identity and visit context
Trust: records, uploads, and family context
- Medical document capture with explicit, understandable steps
- Household and profile patterns where the programme allows
- Policies and settings that stay review-ready for app stores
Support that scales: tickets, FAQs, engagement
- Notifications centre so updates do not disappear in OS banners only
- Self-service answers for the predictable questions
- Help paths aligned to how your teams actually respond
Platform at a glance
One mobile product for iOS and Android: booking, visits, prescriptions, diagnostics, pharmacy, and documents share navigation patterns so patients learn the app once.
Integrated with existing provider services—authenticated sessions, real schedules, traceable actions—not a marketing shell over a desktop-only portal.
Where partner journeys change quickly, embedded flows complement native screens so you are not blocked on a full rebuild for every policy tweak.
Journeys in the product
Screens are grouped around jobs to be done—book, attend, refill, investigate—rather than one undifferentiated menu. Shared list and card patterns keep behaviour predictable as modules roll out.
Booking & intake
- Search doctors, locations, slots
- Confirm and reschedule with clear state
Visits & telehealth
- Scheduled list and detail views
- Video visit entry where offered
Prescriptions
- History and detail from the dashboard
- Path to fulfilment partners
Lab & diagnostics
- Catalogue and detail flows
- Results handoff in-app
Pharmacy
- Browse and detail with imagery
- Tie-outs to ordering where integrated
Records & uploads
- Medical document capture and attachments
- Profile-linked storage patterns
Family & profile
- Household members where enabled
- Edit profile and credentials
Engagement
- Notifications centre
- Information and FAQ surfaces
Differentiation
AI-native EHR & revenue cycle
On this engagement, the patient app was the member-facing surface; the same programme shipped an AI-heavy EHR and revenue cycle behind it so eligibility, coding, denials, and cost clarity could feed APIs the mobile product consumed.
Where ERP meets the clinical and billing front line: intelligent features that remove repetitive data entry, surface denial risk before submission, and explain patient responsibility in plain language—always with clinician or biller confirmation where regulations require it.
Daily workflow automation
AI insurance card scan
Patient or front desk photographs the card; OCR plus structured extraction fills payer name, member ID, group, plan type, copays, and effective dates. Staff confirm in seconds, then eligibility can run automatically—replacing five minutes of error-prone typing.
AI-suggested coding from clinical notes
As physicians document encounters, the system proposes likely ICD-10 and CPT codes with confidence scores. One tap to accept or adjust—reducing under-coding drift and query volume while keeping the provider in control.
Smart denial prediction
Before a claim goes out, models flag probable denial drivers from historical payer behaviour—missing modifiers, code pairs, prior auth gaps—with concrete fixes (for example: “Aetna: add modifier 25 when billing 99214 with this add-on”).
AI patient cost estimator
After eligibility (e.g. 271 responses), rules combine fee schedules and planned procedures to estimate out-of-pocket cost and generate a patient-friendly explanation—supporting transparency and No Surprises Act-style expectations.
AI-prioritised A/R work queue
Open balances are ranked by dollar at risk, filing deadlines, likelihood of collection, and typical payer response times—each line gets a recommended next action so billers attack the highest-impact work first.
AI claim scrubbing before submission
Pre-837 checks catch missing fields, invalid combinations, outdated codes, modifier conflicts, missing authorisations, and duplicate claims—mixing deterministic edits (CCI, age/gender) with model-assisted edge cases.
Strategic intelligence layer
Turns billing from reactive clean-up into proactive revenue optimisation—patterns no single practice sees alone.
Denial pattern intelligence
Continuous analysis of denial history surfaces systemic payer quirks and auto-generates “cheat sheets” per payer—alerting teams before the same mistake repeats and cutting re-bill cycles.
Ambient pre-authorisation intelligence
When a superbill includes certain CPTs, the system infers whether prior auth is likely required, lists documentation payers usually want, and drafts a first-pass auth request from the clinical note—learning which language speeds approvals.
AI call prep for denial follow-up
Before a biller phones the payer, they get a brief: talking points, appeal phrasing that has worked for this denial reason, department routing, expected hold times, and documents to have ready—capturing expertise that usually walks out with senior staff.
Patient payment propensity scoring
Scores likelihood to pay by history, balance, plan type, and context—recommending whether to collect at check-in, text a statement, offer a plan, or flag financial counselling—reducing bad debt and redundant dunning.
Revenue forecasting & cash-flow prediction
30/60/90-day outlooks with confidence bands from pipeline claims, historical adjudication speeds, denial rates, and seasonality—flagging revenue at risk before it becomes a write-off so leaders can staff and plan with CFO-grade visibility.
From assistance → prediction → decision support
AI capabilities are staged by autonomy: start with assistive tasks (capture and coding help), add prediction (denials, payments, cash), then support higher-judgement workflows (auth drafting, appeal calls, queue ranking).
Assistive AI
- Insurance card capture & field mapping
- ICD/CPT suggestions from notes
- Pre-submission claim scrubbing
- Patient responsibility estimates & explanations
Predictive AI
- Denial likelihood & pattern intelligence
- Payment propensity scoring
- Revenue & cash-flow forecasting
Autonomous decision support
- Prior auth drafting from clinical context
- Appeal call briefs and payer-specific guidance
- A/R queue prioritisation & recommended actions
- Cross-tenant alerts as models improve
The learning flywheel
Unlike static rule engines, the platform tightens with volume: more claims improve denial prediction and payer fingerprints; more appeals sharpen call-prep language; more payments refine collection recommendations; more forecasts narrow confidence bands. New sites benefit from accumulated network intelligence while retaining appropriate data boundaries and governance.
How we delivered without losing the clinic’s trust
Releases were sequenced so authentication, appointments, and dashboard stability preceded wider clinical adjacency—reducing the blast radius of API and UX changes while real patients were already in flight.
Follow the real patient journey
We mapped search → book → attend → follow up before polishing edge screens. Appointments, identity, and media contracts were settled early so downstream modules did not thrash.
Ship slices people could rely on
Core booking and dashboard stability landed first; prescriptions, lab, and pharmacy rolled in stages behind shared list/detail patterns—so the product stayed coherent under deadline pressure.
- Consistent error and recovery behaviour across high-traffic flows
- Push and reminders validated on staging before wide rollout
- Permission flows tested against current device and store rules
Production-ready, not “demo-ready”
Crash analytics, upgrade testing across OS versions, and store checklist discipline—so live traffic looked like rehearsal, not a surprise.
Scheduling and fulfilment in one shell
The same navigation home hosts booking, visits, prescriptions, diagnostics, pharmacy, and documents—so patients learn the product once even as the network adds modules.
Scheduling & attendance
- Doctor search and slot selection with confirmation flows
- Scheduled and past appointments with actionable detail
- Video visit entry where the programme supports it
Records & fulfilment paths
- Prescription visibility and pharmacy exploration
- Lab browsing with structured detail screens
- Medical document upload and attachment patterns
Use cases: flows that stay coherent end to end
These patterns describe how screens chain without forcing patients to re-enter the same data at every hop.
Access: search → slot → confirm
- Doctor and location discovery with slot selection
- Confirmation and calendar-friendly success states
- Reminders via push where users opt in
Continuity: visit → record → next step
- Appointment history and detail for accountability
- Prescription visibility to drive adherence
- Hooks into lab and pharmacy without duplicate entry
Trust: documents & identity
- Upload paths for imaging and reports with clear UX
- Profile and family linkage where the programme allows
- Settings and policy surfaces kept in step with stores
Outcomes: what “better access” actually meant
What patients stopped doing
- Calling for tasks the app could complete in two taps
- Re-entering the same details at every hop
- Guessing whether a link in SMS was still valid
What operations could finally measure
- Booking completion and reschedule rates
- Engagement with prescriptions and follow-up surfaces
- Crash-free sessions across OS versions
What we protected on the way
- Runtime permissions aligned with platform policy
- Staging parity before widening rollout
- Bounded embedded journeys where partners move faster than native releases
Where to read more
For ERP and operations positioning for healthcare organisations, see ERP for healthcare companies and ERP for hospitals. All shipped narratives live on case studies.
Healthcare patient app — questions this case study answers
The mobile product was one surface of a broader programme: the provider network also invested in AI-assisted intake, coding, denial prevention, and cost transparency on the revenue-cycle side. That stack feeds APIs the patient app consumes—for example cleaner insurance fields, eligibility-driven cost messaging, and fewer duplicate data-entry steps—so we document both halves in one narrative. See also our healthcare ERP page for the same AI feature set in context.
The team needed one codebase for iOS and Android with predictable async behaviour: sagas isolate login, appointment lists, and other API flows from UI components, so screens stay testable and errors recover gracefully. That pattern scales as modules—lab, pharmacy, documents—ship on staggered timelines.
WebViews are used selectively where embedded journeys or partner content change faster than store release cycles justify rebuilding natively. Core booking, lists, and authenticated API traffic stay in native screens; WebView surfaces are bounded, cookie-aware where needed, and kept out of critical path where stability matters most.
The app uses document picker, image capture, and blob utilities to move files into the provider’s pipeline with explicit user action—no background surprises. Runtime permissions align with current Android and iOS policies for camera, storage, and media access.
They support booking confirmations, reminders, and engagement nudges where users opt in—reducing no-shows and keeping follow-up surfaces visible without spamming generic marketing.
No. This case study is a patient engagement mobile client—scheduling, visits, prescriptions, and adjacencies—integrated with provider services. Hospital ERP typically covers procurement, finance, workforce, and inventory at enterprise scale; see our healthcare ERP pages for that positioning.
Upgrade work included validation against new Android behaviour (for example location and media rules), crash analytics on staging builds, and phased rollout so production matched rehearsal—not a one-day big bang.