I think I accidentally tried to replace my own job with AI
Not even kidding.
I was just trying to improve our API testing setup. You know, the usual stuff — make test generation easier, reduce manual work, maybe use a bit of AI here and there.
Then it slowly escalated.
Instead of just generating tests, I thought:
“what if AI can also understand the API?”
Then:
“ok… what if it can suggest improvements?”
Then:
“what if it can actually implement those improvements?”
At some point I realized I wasn’t building a tool anymore.
I was building something that behaves like a small engineering team.
The setup I ended up with is surprisingly simple:
- one part analyzes the system (what’s missing, what’s risky)
- one part decides what’s the safest way to improve it
- one part actually implements it + tests + checks security
Then it just shows me a demo.
That’s it. I don’t write the tests anymore in that flow.
But here’s the part that hit me.
There was a moment where I paused and thought:
“if this works… what exactly is my role now?”
Because a lot of what I used to do manually:
- writing test cases
- thinking of edge cases
- validating flows
…it’s starting to be handled automatically.
What didn’t go away though:
- deciding what should be built
- knowing if something is actually correct
- catching when something looks right but isn’t
- making sure nothing breaks in weird edge scenarios
- setting the rules so the system doesn’t go crazy
Basically, the responsibility didn’t disappear.
It just shifted.
I’m not really “doing less”.
I’m just… doing different things now.
Less typing.
More thinking.
Still not 100% sure how to feel about it.
Part of it is exciting.
Part of it is a bit uncomfortable, not gonna lie.
Feels like the job is changing in real time while I’m still in it.
Curious if anyone else in QA / automation is experiencing something similar.
Are you still writing most of your tests manually?
Or are you starting to let AI take over parts of it?
Top comments (0)