Release testing for product teams: a guide
Check every deploy like a customer would without a QA department: which journeys to test, where to run them, how to hear about failures and how to keep the suite alive.

The job#
Your team ships often, more of the code is written by AI, and nobody has time to click through the product before each release. Google's 2025 DORA report found that AI adoption continues to have a negative relationship with software delivery stability. The work of catching what broke has to happen somewhere, and it is cheapest right after the deploy, before customers find it.
These posts describe how to do that with plain-language tests that run in a real browser (and on iPhone and Android for apps), with proof for every check.
Decide what to test#
- Release checks for a startup with no QA team
- Turn your release checklist into tests that run themselves
- Check AI-written code before your customers do
- Tests that survive a redesign: why selector tests break
Run them on every change#
- Run browser tests after every deploy with GitHub Actions
- Run plain-language browser tests in GitLab CI
- Run browser tests from any CI with one file
- Test every Vercel preview before you merge
- Run your key journeys against production every night
Test the journeys that matter#
- Test Stripe checkout and plan upgrades in test mode
- Test team invitations from the email to the joined workspace
- Test admin and member roles
- Test receipts, notifications and other emails your app sends
- Test search, filters and sorting
- Check the numbers: totals, prices and counts on screen
- Test that a double click does not create two orders
- Test that a multi-step form keeps answers when users go back
- Check the record was saved, not just the success message
Hear about it, and share it#
- Send test failures to Slack, with the proof attached
- Send test results to Linear or Jira with webhooks
- Share a live test run in a pull request
Keep it trustworthy#
- Flaky tests, and what a test verdict should be allowed to say
- What a browser test suite costs when you pay per run
- Run tests in production without polluting your data
- Test both sides of a feature flag before you flip it
- Upgrade a framework without breaking what users do
- Write tests in your team's own language
The hard parts of the web (iframes, downloads, uploads, drag and drop, slow pages) have their own series with a real run for each: The hardest parts of web testing, each one run for real.
Start this week#
Write the three tests whose failure would hurt most, run them from the app until they pass, then add them to your pipeline. The free trial has 10 runs and needs no card.
Filed underWebReleases and regressionsCI, schedules and agents