đŸ€– Day 3: List ways in which AI is used in testing

30 Days of Testing
, ,
102

AI streamlines testing by integrating comprehensively and providing deeper insights compared to automation. It enables self-healing for immediate issue resolution and simulates tests with created scenarios, incorporating visual validation and learning from interactions.

103

Thanks For the Sharing the information @sarahdeery .

1 Like
104

Después de leer la mayoría de comentarios, me gustaría resaltar las aplicaciones o plataformas que me han resultado mås interesantes. He aquí mi lista:

TestCraft - Me ha sorprendido mucho cĂłmo a partir de un simple click crea a partir de un elemento web casos de prueba con sus respectivos scripts.

Testsigma, Testim, Virtuoso, Apllitools, Katalon y Mabl - Debo explorar todas estas plataformas en profundidad para sacar conclusiones y ver cuĂĄl se adecĂșa mejor a mi forma de testear.

Visily - Muy sorprendido también de lo que es capaz de hacer a partir de una imagen subida.

BlazeMeter - No conocĂ­a esta plataforma para testear renddimiento y tiene muy buena pinta.

105

Hi Testers,

AI can help a lot to improve testing quality . Below are the list of things that can be improved using AI:
-Test Data Generation
-Defect Prediction

  • AI-Powered Automation Testing
  • Optimising Test Code.
  • Bug reporting / Drafting

Tools that can be used are as follows:

  • Code GPT - VS Code Extension
  • Google Gemini
  • Mostly AI
  • Eggplant(Mobile devices)
  • Parasoft Selenic (Integrate with selenium)
1 Like
106

You had already mentioned them all! I’m just starting out, and as a manual tester, I plan to use AI to generate test cases based on requirements and to increase coverage by introducing new tests that I might not have thought of.

1 Like
107

As a part of @parwalrahul prompt activity, I have tried to create a mind-map where I have mentioned the scope of LLM based on my testing activities at the workplace as well as on personal projects, and how LLM can be leveraged for such activities:

Weekly Schedule (4)
Weekly Schedule (4)1704×1002 229 KB

2 Likes
108

You’re doing the prompting for testers course?

2 Likes
109

Yes, I did this task as part of an activity for a prompting course.

2 Likes
110

woah! waiting for you to complete your course and read your feedback!

2 Likes
111

Good day @sarahdeery,

AI in testing is evolving with tremendous speed, giving us a glimmer of how this may alter our perspective on quality. From some of the new tools and use cases that I explored, the following four stood out to me:

  1. Self-Healing Test Scripts
    The tools Testim and Virtuoso, among others, employ AI to improve the positioning of test scripts when changes occur in locators or DOM structures. It feels like an inbuilt test maintainer is looking into releases so that flaky tests do not delay them.

  2. AI-Powered Test Data Generation
    Tonic.ai and Synthesia (for simulation environment) are examples of AI generating synthetic but realistic test data. This would be exceptionally helpful when it comes to privacy-compliant testing, mostly in healthcare or finance domains where real data is not always admissible.

  3. Intelligent Regression Testing
    AI can decide which test cases to execute depending on recent changes in code. Launchable helps in predicting which tests may most probably fail, thereby saving CI time and providing fast feedback loops.

  4. AI Visual Testing
    With help from Diffblue or Applitools, AI-based visual testing continuously analyzes UI layouts to spot even the slightest inconsistencies invisible to human testers. It is very crucial for branding maintenance.

  5. Language to Test Case Conversion in Natural Settings
    Tools such as TestSigma and AutoRegex enable testers (even non-coders) to write test scenarios in plain English, with an AI system converting it into executable test logic. This is making test automation more accessible!

Reflection:
For me, AI-based test prioritization and self-healing scripts would be a total game changer. With fast turnaround sprint cycles, smarter CI builds that truly run only what matters help us practice shift-left without any slowdown. And ideally, automation shouldn’t be a full-time maintenance job: it just should work.

Would love to hear what others are finding! If you’ve tried anything hands-on, do share your experiences.

112

Source: [2409.00411v1] AI-powered test automation tools: A systematic review and empirical evaluation

  1. Self-Healing Tests: Automatically adapts and corrects test scripts in response to system changes, reducing manual maintenance.

  2. Unit Test Generation : Generates unit tests, often integrated into IDEs, enabling rapid test creation for selected code.

  3. Natural Language Processing (NLP): Allows for rapid creation of test scripts using natural language editors and automatically generates test cases/scripts.

  4. Visual Testing Using AI: Leverages computer vision and ML to automatically detect visual and functional regressions by comparing with baseline images.

  5. Unique Features:

  • AI Blueprinting uses ML bots to explore all app paths, generating use cases and verifying existing functionality. Appvance (AI Blueprinting),
  • AI-Powered Test Prioritization uses AI to test critical functionalities first, quickly identifying key problems. The product Aqua (test prioritization),
  • AI Powered Automation interprets and interacts like a user, testing apps securely without source code access. Keysight’s Eggplant Test (AI-powered automation)
  • Live Authoring uses AI to validate tests automatically as you write them, streamlining the authoring process. Virtuoso (Live Authoring)