Case Study · Insurance Brokerage
AI Test Automation Cuts Test Design Effort by 83% for a Leading US Insurance Brokerage
A leading US insurance brokerage firm runs its business on workflows that cross several applications. Testing them meant designing test cases by hand, scripting automation by hand, and sitting through long knowledge-transfer cycles before any of that work could start.
Qualitrix brought in Nogrunt, its autonomous AI platform, to run AI-led quality engineering across the whole testing lifecycle. Nogrunt built context from the client's own knowledge sources, designed the tests, generated Playwright automation, and ran it all with one click. The engagement covered 550 test scenarios, 3,080 test cases and 16,500 test steps. Across the key phases of testing, effort came down by more than 70%.
- Client
- A leading US insurance brokerage firm
- Engagement
- AI-led quality engineering and test automation
- Platform
- Nogrunt autonomous AI platform
- Scope
- 550 test scenarios, 3,080 test cases, 16,500 test steps
01The Challenge
Five constraints slowing down every testing cycle
None of these were unusual on their own. Together, they meant each new round of testing started late and took more effort than it should have.
-
01
Complex workflows across applications
Core business workflows ran across multiple applications. A test had to follow each process from start to finish, not stop where one system ended.
-
02
Manual test design and scripting
Test cases were designed by hand and then turned into automation scripts by hand. Both steps had to finish before a single test could run.
-
03
Long KT and onboarding cycles
Testers had to understand the applications before they could design anything. Knowledge transfer and onboarding took a long time.
-
04
High automation effort and maintenance
Building automation took a lot of effort. Keeping it working as the applications changed added more.
-
05
Dependence on scarce automation specialists
Automation relied on a small pool of specialists who are hard to find, so their availability set the pace for everything else.
02The Approach
Four autonomous stages, from knowledge transfer to execution
Nogrunt moves through the testing lifecycle in order. Each stage feeds the next, so the context gathered at the start carries all the way through to the tests that run at the end.
Autonomous Context Builder
Built from
- KT sessions
- Audio and video recordings
- Documents
- Business discussions
Nogrunt builds a deep understanding of the applications from material the client already had. Testers no longer have to absorb that knowledge before they can contribute. The platform turns it into the application context that every later stage depends on.
Autonomous Test Design
Generated
- 3,080 test cases
- 16,500 test steps
- 91% accuracy
- Complete coverage
Working from that context, the AI generates test cases and the steps inside each one. It produced 3,080 test cases and 16,500 test steps at 91% accuracy with complete coverage. Before, each of those cases would have been designed by hand.
Autonomous Automation Generation
Built-in intelligence
- Reusable components
- Modular design
- Self-healing locators
- Open-source code (Playwright)
Next, the AI creates the test automation assets themselves. Reusable components and a modular design stop the suite from repeating itself. Self-healing locators target the upkeep that usually follows when an application changes. The output is open-source Playwright code rather than a format tied to one vendor.
One-Click Execution
Runs across environments
- Functional tests
- Regression tests
- End-to-end business flows
Tests run across environments with one click. That covers functional and regression tests, plus the end-to-end business flows that span the client's applications. Running a suite is now a single action, not a separate setup job for each environment.
03The Results
Less effort at every stage of the testing lifecycle
Reduction in test design effort
AI-generated test cases and steps replaced manual design across 3,080 test cases and 16,500 test steps.
Resolves 02 Manual test designReduction in automation scripting effort
Generated Playwright assets, built from reusable, modular components, took the place of hand-written scripts.
Resolves 02 and 04 Scripting effortFaster KT and documentation phase
The Context Builder turned existing KT sessions, recordings and documents into application understanding. That shortened the phase before testing could begin.
Resolves 03 KT cyclesFaster environment setup
Getting environments ready to run functional, regression and end-to-end tests took a fraction of the time it used to.
Delivered by Stage 04AI test case accuracy
Generated test cases reached 91% accuracy with complete coverage, so faster design did not mean thinner testing.
Delivered by Stage 02Less reliance on scarce specialists
With test design and scripting generated by AI, test automation no longer waits on a small pool of specialists for every step. Manual effort comes down along with it.
Automation built to be maintained
Self-healing locators and reusable, modular components target the upkeep that made automation expensive to keep running. The open-source Playwright code avoids lock-in to a single vendor.
| Metric | Result |
|---|---|
| Environment setup | 98% faster |
| Test design effort | 83% reduction |
| Automation scripting effort | 75% reduction |
| KT and documentation phase | 72% faster |
| AI test case accuracy | 91% |
| Effort optimization across key testing phases | More than 70% |
| Scope generated and automated | 550 scenarios, 3,080 test cases, 16,500 test steps |
More than 70% effort optimization.
Across the testing lifecycle.
Faster releases, higher quality and significant cost savings for a leading US insurance brokerage.
04The Bigger Picture
What changes when AI does the groundwork
Insurance and financial services teams run on workflows that cross many applications. This client's constraints are common across the sector: long knowledge transfer, tests built by hand, and a reliance on automation specialists who are hard to hire. Each one slows down how quickly a change can be tested and released.
Handing that groundwork to an autonomous platform changes where a QA team spends its time. Less of it goes on manual design and scripting, and less of it waits on scarce specialists. The automation is generated as open-source Playwright code, so the assets are not locked to one vendor. Governance, scalability and security are built in, which makes the platform ready for enterprise environments.
For this brokerage, the result was faster release cycles, higher quality, and more confidence in each release.
Testing complex, multi-application workflows and need test automation that keeps up? Talk to Qualitrix about AI-led quality engineering with Nogrunt.