Earlier this month I sat down with a researcher studying what AI is doing to early-career professionals in the built environment. The conversation was meant to be about graduates, learning on the job and the future of the red pen. Somewhere in the middle of it I said something out loud that I had only been half-thinking, and it has not left me since.
We have spent three years worrying about whether AI is intelligent enough. That question is largely answered. The question nobody is asking is what happens to the human at the end of the pipe when a fleet of agents works through the night and there is ten times more material on the desk at eight o'clock than there was the evening before.
My answer is that the bottleneck moves. It moves from how much we can produce to how much we can organise. And most of us, myself included, are not ready for that.
The maths of the agent era
Think about what a typical consultant's week looked like two years ago. Five live clients, a dozen open threads, a handful of reports in progress. Manageable, just about, with a calendar, an inbox and a to-do list.
Now imagine that agents draft, research, summarise and chase for you around the clock. The work does not shrink. It multiplies. Five clients become fifty, because the cost of serving each one has fallen. A dozen threads become a hundred, because every agent that finishes a task opens two more decisions for you. The intelligence is abundant. Your attention is exactly what it was.
Microsoft's own data already shows the strain before agents have properly arrived. Its 2025 Work Trend Index, based on a survey of 31,000 knowledge workers across 31 markets, reports that 53% of leaders say productivity must increase whilst 80% of the global workforce say they lack the time or energy to do their work. The same report says employees are interrupted every two minutes during the working day by a meeting, email or ping. Microsoft's framing is that we are all about to become an 'agent boss', someone who 'builds, delegates to, and manages agents to amplify their impact'. That is a vendor's phrase and it comes with the usual vendor optimism, but the underlying observation is right. Managing agents is a management job, and management is mostly organisation.
Why a to-do list will not cut it
The tool most professionals rely on to manage their commitments is a list. Sometimes it is an app, often it is a notebook, occasionally it is a set of flagged emails. It works because the number of things on it stays roughly within the range a person can hold in their head.
That range is about to be blown through. The person who cannot see at a glance what they have committed to, what they are waiting on from someone else, and what is parked for later, will drown. They may be every bit as capable as the colleague next to them; they simply cannot find the decision that needs making underneath the material that does not.
I went back to David Allen's Getting Things Done after that interview. I had remembered it as a book from the eighties; it was actually published in 2001, which still puts it before the smartphone and two decades before the first agent. Allen's premise, which his company has since trademarked, is that 'your mind is for having ideas, not holding them'. He was writing about information overload from email and mobile phones. He had no idea how modest that overload would look.
What holds up in Allen's system is the discipline, not the stationery. Every input gets captured in one trusted place. Every input gets clarified into a next action, a thing to wait for, or a thing to park. Every list gets reviewed on a rhythm. Those three habits are precisely the ones an agent boss needs, because an agent is just another source of inputs that arrive faster than a human colleague could produce them.
The skills we keep talking about are downstream of this one
Every panel on the future of work lists the same human skills: judgment, emotional intelligence, creativity, coaching. I have listed them myself, and I still believe in them. But I have come to think that organisation sits underneath all of them.
Judgment requires the time to look properly at what an agent has produced before you put your name to it. Creativity, as anyone who has had a good idea in the shower knows, needs unstructured space that a saturated diary never offers. Coaching a graduate through why an AI-generated report is wrong in this specific scenario is a conversation that has to be scheduled and protected. None of these happen if the person in question is spending their day reacting to whatever surfaced last.
So the person who masters the organising skill does not just get more done. They are the only one who gets to exercise the other skills at all.
The well-being trap
There is a darker version of this story and it needs saying. Several people in that research programme have raised the dopamine hit of a great AI output and the temptation to chase it. I recognise that. I have every tool going and I am working longer hours than I have in years, because the possibilities do not stop and neither, left to my own devices, do I.
If we get the commercial model wrong, this becomes the default for everyone. Firms that use AI purely to cut cost-per-hour will race each other to the bottom, and the efficiency gained will be piled back onto people as more work. Firms that move to outcomes-based pricing, where the client pays for judgment and accountability rather than hours, can pass some of the productivity back to their people as time. Time to think, time to rest, time to be a person rather than an orchestrator of agents who has to stay online so the plates keep spinning.
Organisation is a personal skill, but the conditions that make it possible are set by employers. Companies that want the upside of agents have a responsibility to build the workflows, the routing and the expectations that let people manage the flow rather than be managed by it.
Takeaway
If you lead a team, assume that the volume of work products your people handle will rise by an order of magnitude within a couple of years, and ask whether their current way of tracking commitments could survive that. Almost certainly it cannot.
Pick one trusted system for capturing every commitment, whether it came from a client, a colleague or an agent, and stop keeping anything in your head.
Separate what you must do next from what you are waiting on and what is parked. Agents will fill the 'waiting on' list faster than anything else.
Put a weekly review in the diary and defend it. It is the only point in the week where you see the whole picture.
Treat organisation as a competency you hire for, train for and talk about, in the same breath as judgment and communication.
The firms that get this right will run faster on the same intelligence everyone else has. The ones that do not will have very clever agents and very tired people.
If you want more thinking like this on how AI is reshaping the way built environment professionals actually work, the Project Flux newsletter lands every Monday at projectflux.ai.
Links and Stuff
All content reflects our personal views and is not intended as professional advice or to represent any organisation.
