Episode Details

KPMG deployed an AI recruiting assistant called Kai — built with conversational AI platform Paradox — and after one year, the results are hard to ignore. Interview scheduling time dropped by 58%, from about 60 minutes per candidate down to 25. Across KPMG's US talent team, that adds up to over 1,000 recruiter hours returned to higher-value work. What makes this story stand out isn't just the numbers — it's the philosophy behind them. Kai handles the administrative layer: scheduling interviews, answering candidate queries, surfacing role recommendations. It does not screen candidates, rank CVs, or touch hiring decisions. Every consequential call stays with a human. In year one, Kai handled 23,000 candidate queries — and a full third of those came in after business hours, questions that would have gone unanswered until the next day without the AI layer. KPMG has since expanded the Kai programme to Singapore, showing that this human-plus-AI model transfers across geographies and regulatory environments. The template is clear: AI delivers the clearest ROI in recruiting when it targets the administrative substrate — scheduling, logistics, query handling — rather than the judgment-intensive decisions that define candidate selection. The lesson for any talent leader? You don't need AI to touch hiring decisions to get a significant return. Sometimes the biggest opportunity is hiding in the hour your team spends just getting an interview on the calendar.