Test Maintenance Benchmark
Transparent methodology for measuring governed agent fleets. Results published as available, framework pages labeled clearly when data is in progress. This page documents methodology; results are published when available.
What is measured
Effort to restore green CI after intentional UI and API changes, agent-assisted vs manual baseline.
Maintenance tax destroys ROI of automation; fleets should absorb change-induced breakage.
Methodology
Controlled change sets applied to reference apps; measure human minutes and number of agent interventions to return to pass.
Limitations
Baselines depend on team skill; results report distributions, not single headline numbers, when published.
Request benchmark briefing
Transparent methodology for measuring governed agent fleets. Results published as available, framework pages labeled clearly when data is in progress.
