A team’s weakest working condition, such as constant interruptions or an unclear split of tasks between people and AI, limits its performance. Asking people for more effort or more training does little until leaders find that condition and fix it first, using the HPI framework as the lens.
When People Work Harder and the Team Still Stops Improving
After an AI rollout, the spending often keeps rising. Licenses for AI assistants and a contract with a workflow vendor arrive quarter after quarter, and the team’s output reads about the same as it did before anyone signed anything.
A leader reaches for one of two levers at that point. They ask people to push harder, or they book another round of training. Both answers assume the people are the problem, and both miss where the problem is.
The people on that team are often working as hard as they can. One working condition sets the limit on what they can produce, and the first task is to find it. Effort cannot lift people past a poor working condition, and spreading the budget across many fixes tends to miss the condition that limits the team.
The Weakest Working Condition Sets the Limit
At any given moment, one working condition sets the limit on what people can produce. It might be messages and meetings that break the day into short pieces, so nobody stays with a problem long enough to finish it. It might be an unclear split of tasks, where no one is sure whether the person drafts and AI edits or the other way around, so people do the same work twice.
The founding HPI paper states that performance is set by “the most limiting active constraint,” and that this limit “cannot be sustainably overridden through increased motivation or effort alone.” A motivated team with a fragmented calendar produces about the same as a tired team with the same calendar, because the calendar caps the output. Asking people for more effort at that point adds pressure on them without moving the limit.
Five Fixes at Once Buy Activity While People Stay Stuck
Funding several fixes at once can look like the careful choice. A leader backs five plausible improvements and waits. Only one of them sits on the weakest condition, so the other four buy meetings and a sense of progress while people’s work stays the same.
There is a second cost, which shows up later. When nothing improves, nobody can tell which of the five failed, because all five were running at once. When leaders name the weakest condition first, they can fix it and then see whether people’s work improves. Before more money or more training, someone has to find the condition that limits the team today.
Find and Fix the Weakest Working Condition First
Leaders can start by sitting with the team and walking through how the work gets done. The people doing the work often know where their day breaks down, and they can name the condition that limits them this quarter.
Then leaders fix that one condition and watch how people’s work changes. If it improves, the diagnosis was right. If it does not, leaders test the next one, which is still faster than funding five guesses and learning nothing from any of them.
Fixing the weakest condition brings the next one into view. Protected focus time can reveal an unclear handoff between people and AI, and the work continues from there. Before approving another round of AI spending or another training day, leaders should find the weakest working condition and fix it first, using the HPI framework as the lens.