I’m Bill Echlin. I’ve worked in software testing for far too many years. I still test for a living.
I spent a lot of that time fiddling with selectors and nursing flaky scripts. Hours of nuts-and-bolts work that rarely paid back what it cost. AI can take on that side of the job now. This lets us get back to what matters. And what matters is the process of testing, the quality of it, and being able to scale it. It might even help us keep up with what’s being released. Then again, with development speeding up too, maybe not. I’m living in hope.
I started GTMS because I was struggling to manage AI-driven testing. I also thought I’d learn a lot about AI development and testing along the way.
What I didn’t expect was to learn this more fundamental lesson…
Design and requirement specs go stale because nothing fails when they’re wrong. Tests fail the moment they stop matching the code. So a well-kept bank of test cases and scripts ends up being the most accurate spec of what your code should do. I’ve found that this keeps my AI agents on the straight and narrow while I build GTMS.
I’m starting to see that our test cases really matter as agents write more of the code.
GTMS is tested with GTMS, and I share what I learn on the blog.
If you’re trying this too, I’d like to hear how it’s going.