Hackett is a benchmarking specialist: for years it has measured how the finance, procurement and back-office functions of nearly every big company stack up against best-in-class, and it sells that proprietary comparison data along with advice on how to close the gap. Its AI bet is a sharp argument, that simply buying AI tools does not work, and that the old model of lifting back-office work out and outsourcing it to a labor pool is finished. The fix, in its words, is process-first, not technology-first.

Hackett packages this as the AI XPLR platform, powered by a domain-specific model trained on its benchmark library, and it has pulled IBM in as a joint go-to-market and delivery partner to scale it, alongside process-mining ties to Celonis and the ZBrain orchestration tooling it acquired with LeewayHertz. The platform mines a client's processes, finds where agents actually fit, and designs the workflows, all anchored to Hackett's Digital World-Class benchmark data.

The catch is size and the squeeze in its own model. This is a small-cap firm with roughly flat revenue, and the same AI-platform delivery it is selling can compress the advisory hours that have always been its bread and butter. Net-net, a genuine, well-differentiated niche, it owns the process IP and is right about where the value sits, but it is small and is partly automating its own billable work.