Episode Details

When KPMG — one of the Big Four consulting firms — deployed an AI recruiting assistant called Kai, the goal wasn't to automate hiring. It was to stop wasting experienced recruiters' time on the administrative layer underneath it: scheduling interviews, confirming time zones, answering the same logistics questions on repeat. Year one results were striking. Interview scheduling time dropped 58%, from roughly 60 minutes per candidate down to about 25. That translated to more than 1,000 recruiter hours saved across KPMG's US talent acquisition team. Kai also handled over 23,000 candidate queries automatically — and here's the detail that stood out: one in three of those queries arrived after business hours. Under the old model, those conversations would have sat unanswered until morning. What made the rollout work was the scope. KPMG drew hard lines around what Kai could do. No resume scoring, no shortlisting, no influence over offers or rejections. Every consequential hiring decision stayed with a human. After strong results in the US, KPMG expanded the same model to Singapore — and found it transferred cleanly across a very different hiring market. The takeaway is simple but easy to miss: AI in recruiting doesn't need to touch hiring decisions to transform the function. Target the administrative substrate, keep humans in charge of what actually requires judgment, and the hours you reclaim go back into the work that matters.