This article was originally published by IronSpark People and Business Empowerment Analyst Joe McKendrick on Forbes.com.
Human management may be messy at times, but is AI any better? AI agents don’t posture themselves politically to get ahead. They just do what they’re asked and pump out the work. They can be managed; they can also act as managers. However, will they change the nature of internal politics that shape decisions and interactions in today’s and tomorrow’s organizations? We’ll see.
AI may be ready to serve management roles for certain situations — but the jury is still out on how fair it can be to workers, a recent Wharton study suggested. It seems to work well for ride-share drivers, for example. But for warehouse work that is algorithmically managed, for example, workers are often pushed to their physical and emotional limits without the empathy of a human manager.
In many cases, human managers may find themselves managing fleets of AI agents, though many have expressed a reluctance to do so, especially since there are still many unknowns to this new side of management. Managing autonomous or semi-autonomous agents is still too new of an experience.
But, ultimately, AI managers and agents will be able to operate cleanly without the foibles of human managers, who often get tangled up in political posturing, right?
Consider the twists and turns of human management, as expressed in a recent confessional by one corporate IT director, suggesting what it really takes to run an IT department. Basically, it’s about 1% proficiency and 99% politics.
Tellingly, it’s not about technology and AI – it’s about knowing how to get the best out of people and help them enjoy their jobs. I talked to executives and managers about the pervasiveness of the unwritten rules that will prevail regardless of how much AI is in place.
Here are their words of advice, as well as questions about whether AI agents would be up to performing such roles:
Learn to say no: “One hidden rule is that saying ‘no’ is actually more important than saying ‘yes,’” said Phil Santoro, entrepreneur and co-founder at Wilbur Labs. “In a startup, or any high-growth department, you are hit with constant requests that look like productive work but are actually just noise. We even wrote this into our studio handbook because ruthlessly filtering out distractions is the only way to protect your team’s time for high-impact work.”
Question: Can AI agents learn to just say no? Can they separate value work from the noise?
Praise with purpose: “One inside rule I have learned is that praise, when used thoughtfully, is a powerful leadership tool,” said Tom Thomas, senior data engineering manager at Indeed. “Genuine recognition for work well done builds trust and motivation. As Mark Twain once said, ‘I can live for two months on a good compliment.’” At the same time, leaders must strike the right balance, Thomas added. “Praise should encourage higher standards, not signal that mediocrity is good enough. When recognition is paired with clear expectations for growth, teams stay hungry and aim for more.”
Question: AI agents can say nice things, but can they deliver sincere praise? Can they keep the praise within its proper context for the task or situation?
Be open about budget decisions: “I’ve found that instilling a sense of ownership and shared success consistently pushes teams to think beyond their roles – often resulting in constructive solutions that reduce infrastructure and compute costs,” said Thomas.
Another lesson in budgeting is to encourage “monthly bottoms-up expense reviews to find opportunities to cut overhead without affecting quality,” said Santoro. “We download the statements and go line by line to find software or services that are no longer being used.”
Question: Sure, AI agents can provide alerts or predict if projects or purchases will go over budget. But can they put budget requirements in their proper perspective? Can they redirect funds to other priorities?
Be authentic: “One hidden rule I’ve learned is that authenticity drives growth and performance,” said Gil Pekelman, CEO and co-founder of Atera. “When we started Atera, we asked: is our team having fun, and is our company and team growing? These guide every decision, hire, and behavior. Teams can only thrive if people feel safe bringing their true selves and questions to the table — fear of judgment stifles innovation and progress.”
Question: Would people be comfortable exposing their true selves and questions to an AI agent?
Ignore the hierarchy, build strong teams. “Instead of focusing on the hierarchy, focus on your capabilities,” advised Orla Daly, CIO at Skillsoft. “High‑performing teams rely on skill, judgment, and adaptability. Build the right mix of talent, empower ownership, and foster curiosity and collaboration. Stay close to the work, provide direction, and let experts shape solutions. This is particularly important now in the context of helping team members leverage AI as a tool to create greater impact.”
Question: Can AI agents build effective teams while fostering curiosity and collaboration?
Look for transformation, not just metrics. “Microsoft studied 300,000 employees and found 80% stopped using AI tools after three weeks,” said Jean-Philippe Avelange, CIO at Expereo. “Not because the tools failed, but because there’s a capability gap between what the technology can do and what people can actually accomplish with it. The measurement frameworks we inherit often weren’t designed for the transformation work we’re doing now.” This gap is hard to measure, he added. “’How many people can solve problems they couldn’t solve before’ doesn’t lend itself to easy reporting metrics. So we naturally gravitate toward what’s measurable: logins, clicks, seats filled.
“What’s worked better for me is trying to measure capability alongside activity,” Avelange added. “Looking for signs that people are solving new problems, automating their own work, becoming more effective. It’s messier and harder to quantify, but it’s closer to actual transformation.”