Why AI won't do the hard part of HR ops

SummaryWhat happens when the push to automate HR collides with the humans inside the process? In this episode of The WorkOps Podcast, host Jeet Mukherjee sits down with Nancy Luschkowski, Director of HR Infrastructure & Operations at PagerDuty, to unpack a live onboarding redesign happening amid restructuring, attrition, and the AI wave. Nancy explains why a perfect automation can still ruin the employee experience, why every cross-functional process needs an end-to-end owner, how PagerDuty created a new infrastructure and operations function to keep handoffs clean, and exactly where AI helps (accelerating the starting point) versus where it doesn't (the collaborative hard part). A practical episode for HR ops, people ops, and anyone responsible for employee experience.Chapters00:00 Cold open: the risk of over-indexing on automation01:45 Meet Nancy: why HR operations is the center of everything04:15 The story: onboarding amid reorgs, attrition, and the push to automate06:15 Notification black holes and "enabling the automation"08:45 PagerDuty's new HR Infrastructure & Operations function10:15 End-to-end process owners: one person, the whole experience12:15 Moving fast enough: rebuilding every time an owner leaves13:45 Manual work, tool stack decisions, and source of truth vs. agents18:15 AI as the starting point: process maps, drafts, and the legal persona trick22:45 The ideal onboarding experience: intuitive, customized, human27:45 Measuring onboarding success29:05 Final thoughts: come find me and talk to meTakeaways- You can build the perfect automation, but if the people in it don't know their step or lack context, the experience fails — design automation and the human touch in tandem.- Automation is point-in-time and adoption isn't: managers who hire rarely experience onboarding as brand-new, so one-size-fits-all workflows rarely work and continuous re-enablement is mandatory.- Cross-functional processes need an end-to-end owner — one person accountable for the experience from start to finish, not just a collection of contributors.- Slow down on tool-stack decisions to avoid tech debt: assess current state, define owners, and improve incrementally rather than in one big release.- AI accelerates the starting point — process maps, drafts, persona critiques — but it doesn't do the hard part: incorporating stakeholder feedback, testing, and iterating with a product mindset.Connect with the GuestLinkedIn: https://www.linkedin.com/in/nancy-luschkowski-pmp-shrm-cp-1a955665 Website: https://www.pagerduty.comSponsorThis episode is brought to you by Kinfolk, the AI service desk built for HR.See more at kinfolkhq.com