How TestNova Slashed Test Case Generation Time for an Enterprise NMS
Complex enterprise software demands thorough testing. But manual test case creation is slow, expensive, and hard to scale. TestNova, our AI-powered test automation platform, set out to answer a simple question: can AI generate production-quality test cases directly from UI images, with zero manual annotation? Here is what we found.
The Challenge
Modern network systems are complex. They come with many dashboards, large amounts of data, and thousands of devices to manage. For engineering teams, writing test cases for every screen and every update becomes a never-ending task.
For this study, we worked with a real-world network management system that had:
- A large, always-available design that supports many users
- APIs that allow different systems to connect and work together
- An on-premises setup with strict security rules
- Multiple dashboards showing detailed network and performance data
Even when testing just 4 dashboards, the manual effort was significant. It took about 10 working days and cost around $1,500 to create 245 test cases.
What TestNova Did
The team uploaded 4 NMS dashboard images directly into the TestNova platform. From images alone, the AI identified UI elements, inferred user flows, and generated complete, structured test cases ready for QA.
Four configurations were benchmarked:
- In-House (no RAG): Base model, no retrieval augmentation
- In-House + RAG: Base model with retrieval-augmented generation
- TestNova Model 1: Best overall balance of volume and quality
- TestNova Model 2: Optimized for highest accuracy and precision
Beyond image-based generation, TestNova is built to handle the full QA lifecycle. It can ingest PRDs, Jira tickets, Figma files, and plain text specs, and generate risk-prioritized test plans before every sprint, execute tests across frameworks like Playwright, Selenium, and Cypress, and keep scripts healthy through self-healing automation that fixes itself when the UI changes.
Results at a Glance
TestNova’s AI-powered platform generated production-ready test cases from just 4 dashboard screenshots — no manual annotation, no setup overhead. Here’s what that looked like in practice.
Accuracy and Coverage
TestNova paid models reached close to 90% accuracy and coverage, approaching the manual baseline in quality while running in a fraction of the time.
Quality Issues: Fewer is Better
Irrelevant and duplicate test cases waste QA time. TestNova paid models reduced both to near-zero, dramatically cutting the noise that engineers would otherwise need to filter manually.
Cost and Time
The paid TestNova models complete full test case generation in under 20 minutes at $240 in API costs. That is an 84% cost reduction and 80x speed improvement compared to manual effort.
Choosing the Right Model
Two TestNova paid models stood out. Here is how they compare:
Choose Model 1 when volume matters. It generates 83% of the manual case count at near-equivalent accuracy.
Choose Model 2 when precision is the priority. It delivers the highest accuracy, best coverage, and fewest quality issues.
What This Means for Your Team
No custom setup, no annotated data and no specification parsing. Just 4 images and a platform that knew what to do with them. And that is exactly what makes TestNova different.
As enterprise systems grow and new dashboards ship with every release, test coverage needs to keep up. TestNova’s image-driven approach scales naturally with the system, without adding engineering hours or rising test-writing costs. And with capabilities like predictive analysis, test impact analysis, and a real-time dashboard, teams get full visibility into coverage, risk, and release readiness at every stage. The result is consistent; high-quality test coverage generated in minutes, every time.
That means faster release cycles without sacrificing coverage, less QA time spent on test authoring, cleaner test suites with near-zero duplicates, and a team that scales its testing without scaling its headcount.
