I've been doing QA for over four years. I've shipped apps, caught bugs,
and watched releases go wrong in ways no automated tool predicted.
So when people ask me why I still test manually when AI can generate
checklists in seconds — here's my honest answer.
Because AI wasn't there when the checkout button stopped working
on an iPad in landscape mode. My finger was.
The bugs that actually make it to production
In my experience, the bugs that slip through aren't the obvious ones.
Automated tests catch those.
The ones that hurt are quieter:
A payment flow that completes successfully — but creates a duplicate
order when the network is slow. A push notification that lands correctly
but deep-links to the wrong screen after an OS update. A bug that was
fixed three releases ago, reappearing because nobody re-tested it
when something nearby changed.
I've seen all of these. In real apps. Shipped by teams that thought
they had it covered.
What I actually do before every release
I'm not anti-AI. I use it every day.
But before every release, I still open the app on a real device
and go through the same checklist. Not because I don't trust the
automated tests — but because I've learned what they miss.
A few things I always check on mobile:
- Does the back button behave correctly on every key screen?
- Does the app recover properly when the network drops mid-payment?
- Does the layout hold up on a tablet, rotated to landscape?
- After the last bug fix — did anything nearby quietly break?
None of these are exotic. But they're the ones that show up
in production reviews and one-star app store ratings.
Why I built a template for this
At some point I realized I was running the same checks before every
release — just from memory. So I turned it into a spreadsheet.
If you're shipping a mobile app without a dedicated QA engineer,
you might find it useful:
→ Mobile App QA Checklist — iOS & Android
→ QA Test Management System — for managing test cases and bugs
What's the most unexpected bug you've caught through manual testing
that your automated tests completely missed?
I'm curious — drop it in the comments.
Top comments (0)