AI in QA: Are we automating testing or just automating technical debt?

[weber.st.michael]

Hi everyone! Your comments and real-world insights are very important to me, as the industry is moving so fast right now that official reports rarely capture the full picture of what’s happening in actual production pipelines.

I’m currently rethinking our automation strategy and I keep hitting the same wall: AI agents can generate tests by the hundreds, but the time spent reviewing and debugging them seems to eat up all the promised efficiency.

I was recently reading a deep dive into ai in software testing and it made a very solid point: we need to shift our focus from 'script generation' to 'intelligent orchestration' (self-healing, predictive triage, and risk analysis).

I'd love to get your perspective on this:

1. Do you actually trust AI agents to design the logic of your test cases yet?
2. Or are you strictly using AI for boilerplate tasks like generating mock data?

Looking forward to hearing how you're managing the AI-driven influx in your stack: https://testomat.io/blog/ai-in-software-testing/