Every figure below is an input, including how effective you think a merge gate would be. We supply one number: what ShipGate CI costs. The link updates as you go — share it with stakeholders and they can argue with your assumptions, not ours.
Total software engineers in scope.
Fully loaded cost per engineer (salary + benefits).
% of work hours your engineers spend diagnosing problems in AI-generated code.
Production incidents that required an engineering response in the past 12 months.
Fully loaded cost per incident: eng time + revenue impact + customer comms.
% of those incidents whose root cause sat in code an assistant generated.
We do not have a measured effectiveness figure to give you, and we will not invent one. Set these two yourself — they are the only thing standing between the inputs above and the numbers on the right.
Of the debugging time above, how much do you believe is caught before merge rather than after?
Of the incidents you traced to AI-written code, how many would a blocked pull request have stopped?
Every effectiveness rate above is yours, not ours: ShipGate publishes no measured figure for hours recovered or incidents avoided, so this page asks for your estimate instead of supplying one. The output is arithmetic on your inputs — treat it as a framing device for the conversation, not as evidence.