Every piece of software needs to go through software quality assurance testing. It usually involves a human tester creating a series of test cases, and then checking the software interface for bugs and other issues.
On Tuesday, MuukTest, a Raleigh, North Carolina startup, announced a new AI agent designed to help build those tests in a much more automated fashion.
“Since we started, the vision has been automating software QA technology to reduce test creation to a click, and that has been the whole vision,” company co-founder and CEO Ivan Barajas Vargas told TechCrunch.
These tools enable testers to test every menu, button and operation in the software user interface under multiple conditions to try and catch as many bugs as possible before the software is released to the public.
CTO and co-founder Renan Ugalde says that Barajas Vargas has been involved in software quality assurance testing for 20 years, and they wanted to use Ugalde’s engineering prowess to capture that deep understanding and train an AI agent to help build the test suites.
They have combined a number of AI technologies to put that knowledge to work including multiple large language models, traditional machine learning, computer vision and image recognition. “We trained AI agents to think just like a QA tester, to understand the context within the application – understand what a menu is, what an input is, and when do you expect to see something,” Ugalde said.
This requires reinforcement learning, as well as more information about the overall context and all the experience that the two founders have in QA to translate that into an agent.
Over time, ‘AI agent’ has emerged as a term for AI-fueled software that assists with a task or set of tasks, but to this point there is no standard definition. For MuukTest, it acts as an intelligent assistant, performing some of the more mundane tasks that a human QA tester would have traditionally done.
The founders immigrated to the U.S. from Mexico in 2011. Barajas Vargas eventually settled in at Dell, while Ugalde landed at IBM, as both grew into their careers with increasing levels of responsibility. They joined forces and launched MuukTest in 2019 where the goal has always been to reduce the amount of effort required to generate and run QA tests.
Early versions of their solution used no-code and algorithms to create the tests, but with the new generative AI product, customers can simply describe the kind of test suite they want and MuukTest creates it for them automatically. They can then run the set of tests created by the AI with a single click, greatly reducing the amount of effort required, he said.
MuukTest really began to find product-market fit at the beginning of last year. Even before the AI element was added to the product set, the company was doing well with a 15x revenue increase last year over the prior year, and they think they are ready to grow even faster with the new capabilities.
The company, which participated in Mass Challenge, a Massachusetts-based startup incubator, the year they launched the startup, has raised a total of $6 million between investments and grants. With 36 employees and 10 contractors, Barajas Vargas says the company intends to remain conservative when it comes to spending.
The new AI agent capability is generally available starting today.