Case studies/Wio
WIO needed dependable UI automation across onboarding, identity, credit, and retail journeys without assembling a toolchain. QApilot replaced the stitched-together stack with a single platform built for canvas-rendered apps and banking-specific flows. It now runs the retail regression suite unattended, every night.
·
Results
“Our automation speed is much faster now, test maintenance is far easier, and even a small team can achieve strong automation coverage. As a result, our app quality has improved as we run sanity testing schedulers daily and on demand.”
About the project
Wio is a Central Bank of the UAE-licensed digital bank. Customers open Personal and Business accounts from the mobile app, then save, spend, borrow, and invest in one place. Release cadence depends on proving, every sprint, that onboarding, identity verification, credit, retail, and SME journeys still work on Android and iOS. A small QA group needed one platform they could own end to end — record, save, execute, report, repeat — rather than a set of tools bridged by scripts. Wio website.
Impact
Before
With QApilot
Before
Flutter renders to a canvas. There is no native element tree for conventional locator-based tools to walk, so record-and-playback degrades as soon as the interface changes.
With QApilot
Element-aware recording works against Flutter's rendered widgets rather than an assumed DOM, with execution, reporting, and test-case management in the same place.
Before
Banking specifics compounded the problem: biometric authentication, identity verification, API-driven validation, dynamic test data, and cross-OS execution each defeat a generic tool on its own.
With QApilot
Biometric bypass, Face ID and fingerprint validation, API response reuse, dynamic and random test data, cross-OS runs, and larger-APK support — well beyond UI-only scenarios.
Before
Failed and stopped runs on dynamic app behaviour, with little recovery once a step broke and maintenance compounding as suites grew.
With QApilot
Step- and case-level conditions, assertions, if/else, recovery, exception blocks, and AI healing that re-resolves shifted elements instead of failing the run.
Before
Quality signal arrived after a release decision, with engineers present for long execution windows and little overnight coverage.
With QApilot
The scheduler runs 20.3 hours of test execution per cycle with no engineer present — 97% outside working hours, so squads open a result in the morning instead of starting a run.
Our approach
QApilot joined as a product partner, not a dropped-in tool. The requirement was a single solution a small QA team could own: record, save, execute, report, repeat — on the Flutter journeys that matter for release.
01
Achieving reliable UI automation on WIO's Flutter apps was the starting constraint. Multiple product squads were ramping at the same time, with progress dependent on access, environment, and test-data readiness. QApilot consolidated the stitched toolchain into one platform and moved from exploratory use into sprint-aligned delivery across retail, SME, identity, onboarding, credit-application, and API-heavy flows.
02
The engineering layer covered banking reality: biometric bypass, Face ID and fingerprint validation, API response reuse, dynamic test data, and cross-OS execution. Stability work added conditions, assertions, recovery, and exception blocks. When an element shifts between builds, AI healing re-resolves it. The autonomous crawler proposes coverage so squads extend a baseline instead of authoring every case from zero, and assisted RCA directs time at genuine defects rather than environment noise.
03
A Forward Deployed Engineer model ran war-room support, daily syncs, and one-to-one enablement across squads, plus alignment sessions that bridged documented test cases with actual application behaviour. An extended scheduler now runs daily sanity and on-demand execution: 11,025 steps across 29 test plans and 13 device models, both operating systems, and three application builds on the same day. Step success is 89.3% on executed steps, with skipped steps excluded entirely. Android sits at 92.2% and iOS at 91.3% on the platform reporting basis. Platform enhancements shipped mid-engagement in direct response to WIO's requirements.
Engagement highlights
AI healing re-resolves shifted elements so maintenance falls as suites grow
Autonomous, crawler-led test generation extends an existing baseline
Assisted root-cause analysis on failed and stopped executions
Forward Deployed Engineer war-room, daily syncs, and one-to-one enablement
Nightly scheduler: 11,025 steps, 29 plans, 13 device models, 20.3 hours unattended
Cards, transfers, lending, onboarding, biometrics, Open Finance, and scheduled sanity
What QApilot delivered
Technologies and tools
Biometric bypass, Face ID and fingerprint validation, API response reuse across steps, dynamic and random test data, cross-OS runs, and larger-APK support — well beyond UI-only scenarios.
Step- and case-level conditions, assertions, if/else branching, recovery and fallback paths, and exception blocks, cutting failed and stopped runs toward reliable completions.
Structured automation delivery across retail, SME, identity, onboarding, credit-application, and API-heavy flows, replacing exploratory platform use with sprint-aligned delivery.
Forward Deployed Engineer support during ramp-up, an extended scheduler for daily sanity and on-demand runs, and platform enhancements shipped mid-engagement against WIO's requirements.
Next
WIO needed one platform a small QA team could own on Flutter-rendered banking journeys. QApilot now runs the retail regression suite unattended every night — 11,025 steps, 89.3% step success, and 2.5 engineer-days of execution with no engineer present.
More stories
Geml
Full sanity automation for a pre-launch dating app, in two weeks
Read the story
GrowSari
Regression readiness turned into real automation progress
Read the story
Explore related capabilities
Flutter renders to a canvas without a conventional native element tree, so locator-based record-and-playback broke as soon as the UI changed. Wio also needed banking-specific flows — biometrics, identity, dynamic data, and cross-OS runs — that a stitched toolchain struggled to cover as one system.
QApilot runs Wio’s retail regression suite unattended every night: 11,025 steps across 29 plans and 13 device models, with 89.3% step success on executed steps and 97% of runs outside working hours — about 2.5 engineer-days of execution delivered daily without an engineer present.
Element-aware Flutter recording works against rendered widgets. The engagement added biometric and Face ID validation, API response reuse, dynamic test data, cross-OS execution, recovery blocks, and AI healing when elements shift between builds.
A small QA team owns record, save, execute, and report in one platform, with Forward Deployed Engineer enablement during ramp-up so squads could extend coverage without assembling a new toolchain.
