When AI takes over the tasks, managers take over the relationships
AI agents may be getting smarter, but human managers remain indispensable.
At the Fortune Brainstorm AI conference in San Francisco this month, in a panel hosted by Workday, executives said that the real change is that software is eliminating repetitive work and redefining what good management is: less task supervision and more focus on mentoring, judgment and emotional leadership.
Also read: AI takes over operational work from managers and forces companies to redefine expectations
Stefano Corazza, head of AI research at Canva, said the company’s goal is to build AI around people to give them “superpowers,” not to replace strategic decision-making or the interpersonal skills of managers.
Aashna Kircher, group managing director in Workday’s HR department, said many managers still spend too much time on tedious tasks.
AI agents can eliminate much of this burden, but companies also need to reset expectations, hold managers accountable and train them to make decisions, Kircher explained.
She suggested that companies reflect on questions such as: “What does it mean to be the best coach or team facilitator? What are the skill sets that now need to be developed within teams in an age of AI, where the expectation is judgment, decision-making and creativity?”
Where humans still need to lead
Kate Niederhoffer, chief scientist at BetterUp, distinguished between core, collaborative, and adaptive performance, noting that humans — and especially managers — excel at the collaborative side: alignment, valuing others, and trust between teams.
“And when you start to rely too much on AI or agents to do this work, you see very bad results and also an atrophy of collaboration,” said Niederhoffer.
Empathy and relational support continue to be areas where people significantly outperform machines. If managers offload these tasks to agents, both performance and perceptions of leadership deteriorate, she said.
Danielle Perszyk, cognitive scientist at Amazon AGI SF Lab, argued that managers today are “stuck to a screen”, with productivity tools that end up undermining their own productivity.
She sees AI agents as “universal teammates” capable of handling digital bureaucracy—navigating apps, tracking updates, orchestrating tasks—so that both managers and individual contributors can think more creatively and strategically.
Perszyk hopes teams will spend “a lot less time looking at screens,” but cautioned that current systems only simulate understanding emotions.
His lab is working on “digital world models” and social training — multi-agent environments that mirror workplaces — so that AI can better understand team dynamics and support, rather than replace, the human emotional labor of management.
Toby Roberts, senior vice president of engineering and technology at Zillow, said that as AI absorbs more of the day-to-day wear and tear, managers will have room to focus where human judgment and connection matter most, reshaping discussions about span of control, skills and team design.
