AI can handle things like payroll, product development, onboarding, staffing, and safety seminars to some extent. However, most of us would argue that humanity still plays a role in this type of corporate management. After all, HR professionals operate systems created by humans, and even in an era of massive automation, the human touch is still important.
2026 MIT Conference
I heard a lot about this during a panel discussion at the Imagination in Action event at MIT in April of this year. (Disclaimer: I help run the IIA event every year.) A panel led by Practical.ai’s Michael Hayes looked at some of the ways HR and office work will move forward in light of the emergence of the agent swarm.
In addition to Hayes, the panel included Revelio Labs’ Lisa Simon, Crunchr’s Dirk Jonker and Michael Krebs, and Goldstar AI co-founder and AI consultant Amit Mohindra.
2026 emotions
At the outset, Simon spoke about examining workers in this new economy.
“If you focus on what people are saying, whether positive or negative, skill development can actually be positive,” she said. “So they feel good about their skill development, but the workload and the culture, primarily the workload, is what is driving that feeling down.”
“Everyone is talking about AI, but internal compliance and guidelines are still needed to keep up,” Jonker added. “The real traction is when you see multiple agents working together in a network. I think HR is starting to get there, but we’re basically only half way there, or maybe only halfway there.”
Jonker mentioned “three foundations” and challenges that AI must address: distributed implementation, lack of investment in technology, and untapped capabilities.
“If you look at the people who are leading HR organizations today, not just in headquarters but in the field, if you look at their education, it didn’t include strategy, finance or analytics at all. So we need to catch up a little bit in data technology and data capabilities,” he said.
Mohindra added:
“At least in the past, very few people came into HR to run spreadsheets or build models, let alone build agents,” he said. “So we have a long way to go to make HR more data-driven, more analytical, and more digital.”
Mohindra suggested that there is an opportunity for HR now if leaders are prepared to approach it as an adaptive problem and deal with the ambiguity and uncertainty that exists.
“HR is a very rich environment for all kinds of AI: analytical, generative, agentic, etc.,” he said.
Tackling challenges
Later in the presentation, Krebs explained how swarms can build functionality.
“When you combine a human with an agent or a swarm of agents, it becomes very powerful and can significantly increase the output of a single employee without increasing the number of hours worked,” he said.
He argued that swarms of agents can handle deterministic tasks and atomic units of work, freeing humans up to handle the big picture.
“What it actually does is extract what matters most to the human at the center of the swarm of agents,” he explained of this new use of LLM.
“It’s very carefully programmed to match that person’s output exactly, right?” he said. “To suit their preferences and everything. So it’s not just about typing into chatGPT and getting 10 different responses and being annoyed by the responses.”
leadership and mentoring
“Mentoring and the experience of mentoring people is having a really big moment,” Simon said of the role of leadership in the age of AI.
“We are debating: Will AI replace humans? But certainly, AI has exposed leaders who didn’t really emphasize or try to maximize human potential in the first place. So the traditional form of leadership, this transformational leadership where leaders have a vision and have a compelling vision and they follow people… And adaptive leaders help organizations learn together through uncertainty. That’s the only way that organizations are going to mobilize. To be successful, leaders must help their organizations learn faster than technology advances.”
That’s a mouthful. But now that humanity is at a kind of crossroads with AI, I think it’s appropriate to explain how this form of leadership works.
Jonker added that vision and strengthening vision are of paramount importance.
“If you think about large companies, you have an HR island, a finance island, an operational island, and the real power of AI is to connect these different islands,” he said.
Details about the agent
Returning to the role of herds, Krebs returned to the idea that in tomorrow’s world humans will use LLMs to “not worry about the details.”
“We’re extracting the strategic high-cognitive tasks that are most important to humans because they’re all focused on while the agent is processing more repetitive tasks.
” he said.
But then, as Hayes and others discussed the matter, a realization emerged that this top-level leadership could always be draining humanity.
“The strategic high-cognitive tasks that are most important to humans are extracted because they are all focused on while the agent handles more repetitive tasks.
“Who would have thought that in recent decades, as we spent each day at work thinking about certain difficult things, we would find out that it actually required more mindless tasks to give our brains some kind of load? break during the day. We didn’t think of it as a break, but I feel like we’re learning little by little. ”
In fact, we are all learning in a time when the true values and purpose of human society will be revealed as we battle what some are now calling “God-like AI.”
productivity indicators
I’ll explain it in more detail at the end of the talk. Here, each panelist came up with their own productivity metrics.
“I think the obsession with productivity, the obsession with productivity, is going to come back to haunt us at some point,” Mohindra said. He brought in a valuable indicator of his own, referring to something new called “zeitgeist,” rather than something traditionally spelled out like German “zeitgeist.” In his view, it is a measure of human comfort with AI.
Jonker pivoted to the concept of money, suggesting that stakeholders invest in AI to increase free cash flow, and suggesting that AI be measured as such.
“Personally, I’ve never seen a productivity metric that I think fully captures what we’re really trying to measure,” Simon said in turn, ending with “revenue per person” as a suitable metric.
Krebs suggested token usage per employee as a metric to see how employees are leveraging technology.
This is a useful array of angles, and the talk struck me as a substantive discussion on the adoption of AI in the workplace in general. Certainly now is the time to talk about it. stay tuned.
