Fair point. The team I know using this runs a SaaS platform with around 25 engineers. Before the agent, their QA team was stuck doing triage on weekends after bugs hit prod. Now, the agent blocks PRs that resemble patterns from past bugs—things like changes to concurrency-heavy areas that previously caused memory leaks.

It hasn’t replaced QA, but it shifted their role. Now they spend more time analyzing what the agent flags instead of rerunning test plans. It’s not perfect but it’s made a big difference in stability and team morale.

Also in the process of building: Actory AI https://actoryv3.vercel.app/