The rise of generative AI in software development is happening at breakneck speed. Case in point: by 2028, according to Gartner, a staggering 75% of enterprise engineers will be coding with AI assistance. The question we’re left with is what happens when QA teams can’t keep up the pace? And the answer is AI-powered testing.

Here at mabl, we’ve been at the forefront of AI-powered testing since day one and we understand the urgency of closing the widening gap between AI-driven development and traditional QA methods. That’s why we’re thrilled to introduce two incredible new features to our platform, both built on cutting-edge generative AI technology:

  • GenAI Assertions: Effortlessly validate complex application behavior–including visual elements, AI-powered features, text translations, and chatbots–by using natural language to describe the outcome you’re expecting.
  • GenAI Script Generation: Break down silos between developers, testers, and product managers by empowering anyone on your team to generate code that can cover even the most intricate test scenarios.

GenAI Assertions: Defining Correctness Dynamically

Traditional test assertions, while essential, have a fundamental limitation when it comes to testing AI-generated elements: they rely on static, pre-defined criteria, like verifying a CSS class by its stated name or confirming what the expected inner text is in an element. Once you introduce AI into the equation, you’ll need to verify things that are increasingly dynamic and unique. Defining correctness at this point is a bit more challenging. For example, imagine trying to test the output of an AI-powered chatbot whose responses are never identical, but should always be correct; or verifying that a generated image aligns with what the user thought they should be seeing. Traditional assertions aren’t equipped to handle that level of variability.

mabl’s new GenAI Assertions are a paradigm shift in test validation. Testers use natural language prompts to express what they expect to see, in the context of their specific application. By leveraging generative AI, mabl evaluates the output of the application against that prompt, determining pass or fail based on the contextual understanding. This flexibility unlocks a wealth of innovative use cases, like:

  • Image Verification: Ensure that generated images match expectations. For example: “Verify the image contains bicycles”, “Check the tag number within this photo”, and “Ensure the watermark is properly added”.
  • Text Analysis: Validate the correctness of translations, verify the structure and tone of AI-generated content, and even confirm the proper sorting of items in a table.
  • AI Chatbot Testing: Confirm the logic and relevance of chatbot responses, checking for appropriate content length, formatting, and even inclusions of specific action buttons.

Screenshot GIF of a user providing a natural language assertion to ask if an AI-generated image is correct.
mabl's GenAI  Assertions enables validating complex application behavior, such as confirming the logic and relevance of an AI chatbot response.

The potential doesn’t stop there. As more companies embrace generative AI to personalize user experiences and build unique features, mabl’s GenAI Assertions ensure that those innovative features you’re building are reliable and meet your team’s quality demands. It’s not just about testing what is displayed, it’s also making sure it aligns with the intended purpose.

GenAI Script Generation: Democratizing Powerful Test Scenarios

Traditionally, handling complex test scenarios requires writing intricate code snippets–a task reserved for more technical testers and developers. This creates a barrier for non-coders, bottlenecking test creation and limiting who can contribute to QA processes.

mabl’s GenAI Script Generation dramatically simplifies this process, breaking down barriers and democratizing access to handling complex scenarios. Now, anyone on your team, regardless of coding knowledge, can leverage mabl’s user-friendly interface to effortlessly generate robust test scripts in real-time, trying out the script and making iterative changes based on their needs. In addition to empowering anyone on your team to cover these complex scenarios, these snippets can be added to a centralized repository for reuse across your testing, increasing both efficiency and cohesion. 

Here’s how it works:

  1. Describe your Intent: Simply explain in natural language what you want the script to accomplish. For example, “Return a random date of birth for someone between the ages of 30 and 65 in the format MM/DD/YYYY”
  2. Real-Time Code Generation: mabl’s genAI engine instantly generates the code that performs the action requested.
  3. Test and Refine: Try out the generated code directly within mabl to make sure it performs the action correctly, making any additional changes you need.
  4. Share and Reuse: Save the script for yourself or add it to a centralized library that’s accessible to your entire team.

Examples where script generation can help include:

  • Generating unique strings or values that meet specific criteria (e.g. simulated product ID codes, email addresses, names, or geographic locations)
  • Counting the elements on the page that meet certain criteria (e.g. how many of these boxes are checked or how many items are in a list)
  • Interacting with the browser in a specific way (e.g. scrolling up or down the current page)
  • Extracting and manipulating specific information from the URL (e.g. extracting a form ID from the URL to save as a variable for later use)

Screenshot GIF of a user using natural language to describe what they want to do and the mabl AI script generator providing the code to do that task.mabl's GenAI Script Generation empowers anyone to generate code that handles complex test scenarios.

The mabl Difference

With the hype of generative AI, many organizations are “AI-washing” their products wherever they can, simply to garner attention in the market. Here at mabl, AI has always been a part of what we do. For the last seven years, we’ve built our product based on applying AI in the right places with the right use cases to solve real problems, improving the lives of quality practitioners along the way. From automating testing for use cases that once could only be accomplished manually, to opening up quality engineering to the entire team, we believe that augmenting low-code testing with AI as part of a unified platform presents the best opportunity for quality to keep pace with software development.

The Future of AI-Powered Testing

There is no doubt that the generative AI space is rapidly changing. You’ll continue seeing mabl incorporate powerful AI features into the most useful and business-critical areas, ensuring you can quickly and easily create, maintain, and scale robust automated tests. The bar for "easy" is constantly being raised, and we're leading that lift. 

We’d love to hear what challenges AI has brought to your organization and invite you to take a look at our newest features to learn how we can help tackle those challenges. While script generation is available to all users, GenAI assertions are currently available for all mabl customers as an early access program. If you're on a trial and would like to test out these features, please reach out to the mabl team.

See how mabl applies AI to the entire testing lifecycle by registering for a 14-day free trial