Did your AI spending pay off?
Most AI business cases are judged on usage and on what staff say they saved. Neither tells you what changed. We measure the effect the way an evaluation would: against a fair comparison, with costs counted and the uncertainty stated.
Questions this answers
- Did the tools change output, speed, quality or cost, and by how much?
- Which teams and tasks gained, and which did not?
- Does the vendor’s business case hold up when it is rebuilt from your own data?
- Should the rollout grow, change or stop?
Methods
A baseline before rollout, a staggered rollout or randomised pilot where one is possible, difference-in-differences and event-study designs where it is not, full cost accounting including review time, and a return figure with its range. Self-reports are measured, not trusted: in one randomised trial, experienced developers were 19% slower with AI tools while believing they were faster.
Other services
Programme and policy evaluation
Find out whether a programme changed outcomes, for whom, and by how much.
Pay-gap and pay-equity analysis
A defensible pay-gap analysis for reporting rules, and a clear view of which gaps are explained and which are not.
Health economics
Cost-effectiveness, budget impact and model audits that hold up in front of reviewers.