Generative AI-Powered Visual Validation
Relicx visual validation uses Generative AI to completely automate the visual regression testing.
- Laborious manual baselining: Visual validation typically relies on comparing full or partial screenshots with a manually created baseline. Creating and maintaining these baselines can be extremely painful and time-consuming.
- Noisy diffs: Traditional visual validation techniques rely on pixel diffs. This results in noisy diffs and false alarms because of the inability of these tools to differentiate between real visual regression and expected data/content differences.
Relicx leverages Generative AI to address these challenges.
Relicx uses Generative AI to analyze the application page structure and identify the key components (e.g., navigation bar, headers, body, etc.) that make up this page. The ability to identify and diff visually at the component level ensures that Relicx can detect visual regressions with minimal noise.
Relicx AI analyzes the visual diffs to ascertain if the changes are content-related, visual or style regressions, or functional bugs.
Visual Regression Example: Below is an example of Visual Regression detected due to missing icons.
Functional Diff Example: Here is another example showing a functional difference discovered visually. Please note that the "Banking Accounts" and "Transaction/Transfer" category headings are missing in the navigation bar. Such regressions are extremely hard to detect using tranditional testing techniques
Visual validation is automatically triggered as a part Automatic Release Validation.
- Click on the "Validate Release" button to start a end-to-end release readiness validation, which includes visual regression testing
- Once the validation run is completed, the flow map will be color coded to identify the pages with visual issues
-
