How a Global Bank Moved ~1,000 Test Scripts from a Proprietary Tool to Playwright

~81%

Of Standard Tests Converted Automatically

25%

Less Regression Time After Migration

$0

Proprietary License Cost After Migration

We expected a migration project. What we got was our own test suite handed back to us as code we could read, version and run on HyperExecute with our own CI/CD.

Head of Quality Engineering

Digital banking, retail customer channels, Global bank

A global bank migrating its customer-website test scripts to Playwright with TestMu AI

Overview

A global bank moved ~1,000 customer-website test scripts from a proprietary tool to Playwright, cutting license cost and regression time in one move.

The migration used the TestMu AI Framework Migration Agent and HyperExecute. The only input was one successful run of each test on TestMu AI. The output was Playwright code the client owns, traceable to evidence, and running with CI/CD on HyperExecute.

The Problem: Years of Regression Tests Locked in a Proprietary Tool

Years of regression tests for the customer-facing banking website lived inside a proprietary tool and could not leave it. There was no readable code, no Git, and no other runner.

License spend kept growing, while the only exit on offer was a months-long manual rewrite of login, payments and account journeys. That rewrite carried its own risk: every hand-written script could drift from what the original test actually checked.

What TestMu AI Did

TestMu AI asked for nothing but a run. Each test ran once on TestMu AI, and that run was the input. No source assets and no proprietary workspace were needed, which made for a shorter security review.

Playwright scripts were then generated from the run's logs and video, so every step is traceable to evidence. Unclear steps were flagged, not guessed.

Finally, ownership was handed over. The client owns the Playwright code and runs it with CI/CD on TestMu AI HyperExecute.

How It Ran: Four Streams, All in Parallel

Stream 01

Run as-is

Existing proprietary tool tests executed on TestMu AI, unchanged.

Stream 02

Capture

Logs and video recorded per run. A short pilot sized the full job.

Stream 03

Convert

Framework Migration Agent produced Playwright code, plain-English steps and a migration report per build.

Stream 04

Run at scale

Converted tests went live on HyperExecute, batch by batch.

~1,000

Test scripts migrated

~50 / day

Simple and complex scripts converted per day

~4 wks

End to end, streams running simultaneously

Conversion Rate by Test Type

~81%

Standard tests (up to 20 steps) converted automatically

~70%

Complex tests (20+ steps, multiple integrations) converted automatically

The remainder was listed in the pilot report with reasons, not forced through. Out of scope: tests that depend on OTP or live card authentication, expired test data or unreachable environments; desktop and API tests.

One Test, Before and After

A customer "Add payee and transfer funds" journey on the banking website, recorded once in the proprietary tool and converted.

Before · Proprietary tool execution

261 steps

54 seconds. Dense, tool-specific, unreadable outside the proprietary tool.

After · Playwright on HyperExecute

32 steps

41 seconds, same journey: 25% faster, readable by any engineer, 8 tests running in parallel.

Business Outcomes

The bank exited the proprietary tool without a rewrite. ~1,000 scripts were migrated with no line-by-line re-authoring and no engineer months spent on it. The client owns the code, so there is no vendor lock-in.

Regression is 25% faster. The suite runs in parallel on HyperExecute, so release cycles shorten without adding headcount.

The suite is also audit-ready by design. Every converted step traces back to the original run's logs and video, ready for internal audit and regulator review. And the documentation came for free: plain-English test steps now serve manual QA, compliance and new joiners.

The Takeaway: Run It Once. Own It Forever.

The only input is one successful run of each test on TestMu AI. The output is Playwright code the client owns, traceable to evidence, running with CI/CD on HyperExecute. No rewrite, no lock-in.

Ready to improve your testing? Schedule a demo to see how TestMu AI can help your team achieve similar results.

Client

Global bank, retail customer website

Migration

Proprietary tool → Playwright

Scale

~1,000 web test scripts

TestMu AI Products used

More related customer stories

left arrowright arrow

TestMu AI for Enterprise

Get access to solutions built on enterprise-grade security, privacy, & compliance