According to a recent LinkedIn post from mabl, the company made its performance-test gates blocking for all pull requests, including changes submitted by its co-founder. The post says the policy followed a bug in mabl’s own performance-testing feature that was identified before a change could be merged.
The post links the change to the growth of AI-assisted development, which it says increased weekly code changes more than fivefold to roughly 300. It also describes an earlier process in which five employees maintained 70% to 80% of the test suite, contributing to test drift, staff strain and a tendency to overlook recurring failures.
According to the account, mabl’s impact-analysis gate identified three bugs before merge during its first four weeks and flagged three additional tests for updates. The operating rules include tracking emergency bypasses, setting procedures for third-party outages, removing unreliable tests from blocking gates and providing a support channel for blocked contributors.
For investors, the example illustrates how software-testing providers may use their own platforms to manage the higher release volume associated with AI-assisted coding. The results could support mabl’s credibility in quality automation, although the post reflects an internal experience and does not by itself establish broader financial benefits, customer adoption or revenue growth.