Your Team Got Faster Tools And Less Slack
Efficiency did not hand the week back. It raised the complexity ceiling, and the bill arrives in your absence data before it arrives in your output.
Every tool bought in the last three years removed effort from a process. None of them reduced what your teams are carrying. There are five well documented reasons for that, and each one has an operational consequence you can act on. Worth reading before the next resilience workshop gets commissioned.
What is going on
Jevons Paradox
The cheaper a process becomes, the more of it the organisation commissions.
First observed in coal consumption in the 1860s and reliably true of knowledge work since. When a task gets cheaper in effort, total demand for it rises rather than falls, because the constraint that was rationing it has gone.
→Plan for demand to rise when you remove effort. Unplanned, that rise lands on individuals as unpaid overhead.
Work Expands To Fill Cognitive Bandwidth
When one layer of difficulty disappears, the ceiling rises rather than the floor.
Parkinson’s Law applied to attention rather than to time. Removing the straightforward part of a role does not create breathing room inside it. It hands the same person a harder problem, usually without a rewrite of the job description or the grade.
→Automation promotes people. Job design, workload models and grading rarely get rewritten to match, so the promotion is invisible except in how tired everyone is.
Cognitive Load Shifting
Automation changes the type of thinking required, not the amount.
When a tool takes over execution, the human keeps designing the system, checking the output, catching what the tool missed, and switching between the six places the work now lives. That is a different kind of load, and a more expensive one.
→The load moved from execution to orchestration. Orchestration runs on executive function, which is exactly the resource that fluctuates in ADHD, autism and chronic illness.
The Automation Paradox
The more you automate, the more demanding the remaining human role becomes.
Machines absorb the predictable work. What is left for people is ambiguity: the edge cases, the judgement calls, the integration between systems, the conversations that go wrong. The residue of automation is the hardest part of the original job.
→You have concentrated the hardest work into fewer hands. Check whether the support around those hands was concentrated too.
Complexity Inflation
Tools cut the effort per task. Expectations rise to absorb the saving.
Your competitors bought the same tools in the same quarter, so the standard lifts for everyone at once. Output per person rises and the baseline recalibrates upward behind it, which is why the saving never shows up as time.
→The sense of running to stand still is a structural fact about the work. A structural fact needs a structural response, which is why resilience training reads as an insult to the people receiving it.
What to do about it
What to do about it
Treat cognitive load as a planning input, the way you already treat headcount.
The binding constraint on your teams is no longer hours. It is switching cost, decision density, and the recovery time between high context pieces of work. All three are measurable and plannable. They are also the specific things your neurodivergent and chronically ill staff cannot absorb quietly, which is why they reach the limit first and why they are the most useful early warning system you have.
The reframe
"We need this team to be more productive"
"We need to manage how much high context work sits in one person’s week"
Work, admin, emotional labour and life all draw on one executive function budget. Adding to any of them spends the same account.
Recovery time between demanding pieces of work is a scheduling decision rather than a wellbeing nicety. It determines whether the next piece lands well.
Capacity aware planning is an operational control. Choosing the rhythm of a week deliberately is a management skill, and it can be taught.
The business case
Efficiency raises demand, removing easy work promotes people into harder work, and expectations recalibrate to absorb the gains. None of that is a performance problem, and none of it responds to a resilience intervention. It responds to job design, workload modelling, and a week planned around real capacity.
If you want one number to watch, take the gap between how much high context work a role holds now and how much it held two years ago, alongside short notice absence in the same teams. If both have moved in the same direction, you are looking at one cause twice.