From Flaky to Fixed: A Design Thinking Framework for Common Test Automation Challenges

It is 2025, and we are still struggling to find the right set of tools and technique to overcome common test automation problems, teams across globe face. We changed our tool set, our strategies only to realize that sooner or later we come across the same pitfalls. We fail to address the root cause of misery and keep thinking that some divine tool out there exists which will take away all the challenges of test automation, like the bugs of the product we are testing, and all be well in this world. Reality strikes hard when yet another time the strategy to fix all the problems with the magic AI solution also ends in a disaster, this time maybe we might want to go back to the board and ask Why?

The idea to apply Design Thinking principles to ideate to build better products isn’t something new. It has been always there, but maybe we as Test Automation developers and architects failed to see Test Automation as a product, as a development project. Even today, just like testing, test automation is an afterthought. And with that comes the struggles for the team to build the perfect connotation which will identify all issues build after build without itself needing any support or help. This talk aspires to help achieve a shift in mindset of teams enthusiastic to explore, adapt and build strong test automation solutions and be open to ask Why, when things don’t end up as they should.

Design thinking offers a human-centered approach to problem-solving that emphasizes understanding root causes before generating solutions. When applied to test automation, it shifts our focus from fixing broken tests to understanding why our automation approach is failing. In this talk we will explore few of the most common problems we face when building test automation solutions, and how we can use the principles of Design Thinking which are – Empathize, Define, Ideate, Prototype and Test. We will explore solutions like 5Whys and Conflict of Interest to explore possible root causes of problems commonly faced and how we can resolve them.

Having said that, we need to understand that there exists no one solution for designing test automation project. A lot depends on what exactly is the problem, we are trying to solve, what resources we have, what is our ability to build that capacity and what solutions already exists in the market. It is also time to understand and accept that no one solution fits all bills. For some projects maybe record replay approach of scripts works well, whereas for another project, more code architecture and framework design of approach is required, whereas for some solutions of AI agents are more viable.