Blog by : Karine St-Onge
🌍 This article is also available in French: Mon virage émotionnel avec l’IA – Karine St-Onge
When AI barged into my daily life
AI first burst into my life the way it does for many testing specialists: through LinkedIn, conferences, a bit of monitoring, and a lot of noise. Everyone was talking about it with impressive certainty, as if the future of QA had been downloaded overnight.
I told myself: if I don’t get into it now, I’m going to end up like a framework that’s never been updated: functional… but obsolete.
So I started to explore.
The first attempt: disappointment without the drama
An ordinary workday, I try out the idea of the moment: generating a list of test cases from a functional analysis. No precise plan, no context, no structure.
The AI spat out a list of tests that looked pretty… but disappointing at the first serious glance. Gaps everywhere, half-baked ideas… I would have had to reread, verify, complete. I might as well redo the whole thing!
I caught myself shrugging: “Well… it takes longer to supervise the AI than to do it myself.”
I wasn’t angry. It wasn’t a drama. Just a simple but genuine disappointment. Like biting into a cookie that looked better in the photo.
The mirror I hadn’t planned on seeing myself in
As I went back over my attempts, I understood something that isn’t very pleasant at first: The problem wasn’t the AI. The problem was my requests.
I was giving it a ticket with no context, three lines of intent, zero business constraints, and I was expecting magic. It was like asking a first-day intern to write a complete test strategy by just saying: “Here, here’s a Jira. Figure it out.”
By changing my approach, I first asked it to ask me questions in order to build a better context for it, and the result changed immediately.
And that mirror is very revealing: when the AI didn’t know… it was because I hadn’t specified it.
This realization wasn’t comfortable, but it was clear-eyed, almost soothing.
A first glimmer of hope: it’s not impossible… it’s just demanding.
The first real success: the positive jolt
The first few times the AI really helped, I had a real moment of astonishment.
Something simple: a small error in the interface that I hadn’t seen. Then a forgotten related test case, then a silent side effect.
It was nothing spectacular, but it was enough to trigger a physical reaction: I sat up straight, blinked, and reread it a second time.
And in my head: “Okay, now… something’s happening here.”
For the first time, I was no longer alone in watching over the system. I had a second pair of eyes. A pair that never gets tired.
The bittersweet guilt of letting go of a little weight
From then on, the AI became an essential step in my workflow: diff analysis, gathering information in Jira, suggestions of points to watch, ideas for additional tests.
And there, I started to feel two things at once:
1. An immense relief. No longer carrying everything alone.
2. A twinge of guilt. Because it was pleasant. Too pleasant.
I could hear myself thinking: “Am I allowed to be relieved by a machine?”
And the answer, in the end, was clearly yes.
The QA role that regains its color
Because the AI reminded me of a simple truth: The AI reads, but I understand.
The AI proposes, but I judge.
The AI detects, but I analyze.
And when it gets carried away flagging a false problem, I’m the one who reassures. When it misses a real anomaly… I’m the one who catches it.
QAs don’t disappear, they shift perspective. We don’t become executors, we become conductors.
Creativity waking from its torpor
And then, something I hadn’t seen coming: my creativity came back.
Through sheer routine, I had frozen up a bit in my methods. Not out of boredom, just from the steadiness of the routine.
The AI opened a door in my mind: Looker Studio, which I had never had the courage to explore. Apps Script in Google Sheets. Setting up a new testing framework.
One day, I realized my brain was buzzing. It had been a long time…
When the machine eases the loneliness
I’ve always worked alone, carrying almost everything.
And the AI filled a space I hadn’t named: technical brainstorming, a logical mirror, sorting ideas, rationalizing emotions.
In the middle of a difficult stretch, where the team kept looping over the same bugs without managing to resolve them, I talked with the AI. Objectively and calmly.
And the inner overload settled down, like a hand placed on the shoulder. Without judgment, without a sigh, without a “you should have.”
That conversation allowed me to propose a new strategy to the team with confidence. Not because the AI had decided, but because it had helped me clarify my thinking.
If I had to take away one thing
I lived this journey as an emotional crossing: distrust, disappointment, clarity, wonder, relief, reconnecting with my worth… and in the end, a little less loneliness.
I hope that, somewhere, you’ll recognize yourself. Whatever point you’re at with AI, your emotion is legitimate.
It’s not a neutral tool, like a screwdriver. It touches our expertise, our ego, our security.
And it’s normal for it to stir things up.
If my journey can help you think: “I’m not alone in what I feel,” then the goal is achieved.