Our company built a massive logistics platform with more than 100 different screens and modules. The thing is, we never wrote any automated frontend tests at all. Our workflow goes like this: developers finish coding a feature, business analysts check if it meets requirements, then our QA team goes through everything manually and finds basically every bug that exists.
The QA people are really good at their job and catch issues we would never think of. Because of this, I never thought we needed automated testing. But now I’m wondering if we’re missing something important.
Part of me thinks we should add frontend testing, but I can’t figure out why exactly. Teaching our developers how to write tests would take a lot of time since none of them know how to do it. The features we build can get pretty complicated, but our current process works well. We code, do some quick manual checks, send it to QA, then fix whatever they find.
Am I wrong to think this way? What are your thoughts on this situation?
Interesting situation! What happens when your QA team goes on vacation or gets buried with other projects? Also, how long does a full manual testing cycle take for 100+ screens? Your team sounds solid, but I’m curious about timing.
The real payoff comes during maintenance and scaling. Sure, your current process works with dedicated QA, but automated UI tests aren’t meant to replace manual testing. They’re safety nets for when you’re refactoring code, updating dependencies, or pushing hotfixes - times when you might skip manual testing because of deadlines. With 100+ screens, you’ve got critical user flows that could break from changes that seem completely unrelated. Start small - just cover the must-have paths instead of trying to test everything. Training developers upfront is worth it when you’re pushing an urgent fix at 2 AM and need to know you didn’t break basic functionality.
you’re missing regression testing tho. QA catches new bugs, but what happens when you deploy and break something that worked before? automated tests catch that stuff instantly - you can’t rely on QA to retest everything every time.