Applying Tech Innovations in People & Organization: Balancing Automation and Human Interaction
Poor AI
The role of technology in People & Organization management gets debated endlessly, and I have been on both sides of it. HR teams are staffed lean, the push to innovate is loud, and AI agents and self-service platforms are going in at 100 mph. The question worth asking: is technology making P&O more efficient, or is it removing the part of the job that actually matters?
Lean staffing creates a natural pull toward automation, and I understand it. Self-service portals answer the routine questions. Agents go further and close the cases: leave requests, verification letters, benefits changes, onboarding steps, done end to end without a person in the loop. That part works, and I am not going to pretend otherwise.
The cost shows up elsewhere. An agent that resolves a leave request end to end does not just answer a question, it removes the entire interaction. No intake call, no follow-up, no moment where someone mentions their manager has been impossible since March. Those exchanges were never really about leave policy. They were how HR found out what was going on. Automate enough of them and the function goes quiet, then compensates by sending engagement surveys to learn what it used to pick up by walking around. Without a trusting relationship with the people behind the survey, employees are less likely to say anything honest.
I recently talked to a Workday analyst who spent months getting a new feature configured and launched. It went live and then it disappeared, because nobody at the company actually needed it. No use case. Months of work into something that was never going to be used, and this happens more than anyone admits. The operational side of HR decides something will be useful, then does not have the bandwidth to go talk to people and check. That is the same bandwidth problem, showing up as wasted implementation instead of missed signal.
So the target for automation should be the mundane operational layer, not the interaction layer. When the team is not juggling spreadsheets and slide decks, it has time to understand what individual employees are actually good at, what they want, and where they are headed.
The bigger constraint is staffing ratios. A People Partner covering 500 employees cannot build meaningful relationships with 500 employees, no matter how much casework gets automated away. Automation buys hours. It does not buy attention. If the goal is depth of engagement, the ratio has to move too, and that is a budget conversation, not a technology one.
HR professionals are not just administrators. They are counselors, mentors, and strategists, and none of that scales through a portal. I do not have a clean rule for which tasks belong to agents. Anyone selling you one is guessing. What I do know is that the question worth asking is not how much can be automated, but which tasks lose their value when the human is taken out of them. I am still working out where that line sits in my own function.
