Hackett’s new report on AI in HR: read the “up to” first

The Hackett Group put out a short report in June on AI across the HR lifecycle (link in comments). The headline numbers are big: up to 61% lower recruiting cost per hire, 119% higher recruiter productivity, 57% faster time-to-fill, 52% less admin time for HRBPs.

Before you put these in a deck: every figure is “up to” and comes from Hackett’s own modeling of a best-case scenario against a peer median. Hackett sells that benchmarking. Treat the numbers as direction.

The argument is the useful part of it. AI added to single tasks (screening, scheduling, an HR chatbot) gives you local gains that stay local. Larger gains come from redesigning the whole process. Their one case: a healthcare organisation rebuilt its recruiting process, cut time-to-fill from 65 to 35 days.

Something to try this week

Pick one process you own from start to finish: hiring for one role family, onboarding, or internal moves.

1. Map every step, including the handoffs and the chasing.

2. Mark where AI (Claude, ChatGPT, Gemini, Copilot, whatever you use) touches it today.

3. Ask: if I designed this from zero with these tools available, which steps would still exist?

Step 3 is where most stop and go back to speeding up step 4 of 12. Stay with it for 20 minutes.

Where in your HR work are you using AI on one step while everything around it stays manual?