Summary. The four-box matrix that separates the urgent from the important was first published on this page as a productivity tip during the pandemic. The tool has not aged, but the evidence behind it and the way we use it have changed. We clarify three things. First, the famous line attributed to Eisenhower is not his: in his 1954 speech he quotes an unnamed former college president. Second, the “mere urgency effect” experiments show that people pick a task simply because its window looks short, even when they know it pays less. Without a payoff reminder, the share of people choosing the lower-paying task was 18 to 41 percentage points higher in the urgency condition than in the control group. Third, in 2026 the new owner of the “delegate” box is often an AI agent. In Stanford’s 844-task study, workers were positive about automation for 46.1 percent of tasks. Below is a CEOtudent framework that turns the matrix into human, agent, automation and delete decisions.
The real source of the line: 19 August 1954, Evanston
Dwight D. Eisenhower, who gave the matrix its name, spoke on 19 August 1954 on the Northwestern University campus, at the Second Assembly of the World Council of Churches. According to the text in The American Presidency Project archive, Eisenhower quotes “a former college president”: “I have two kinds of problems, the urgent and the important.” The quote goes on to say that the urgent are not important and the important are never urgent.
Eisenhower calls this dilemma a problem of modern man and hopes the assembly can give the important “the touch of urgency.” There is no four-box diagram in the speech. What we now call the “Eisenhower matrix” is not a chart Eisenhower drew; it is a way of thinking that later took his name.
The person who popularized the matrix was Stephen R. Covey. In The 7 Habits of Highly Effective People, published by Simon & Schuster in 1989, Habit 3 is “Put First Things First.” FranklinCovey summarizes the habit as organizing and executing around your most important priorities. The same source says the best return on your time comes from spending it on things that are not urgent but are important. The classic four boxes:
- Urgent and important: Do it now.
- Important, not urgent: Schedule it and protect it.
- Urgent, not important: Delegate it.
- Neither urgent nor important: Delete it.
What the research says: the “mere urgency effect”
Is the problem the matrix assumes real? A 2018 study by Meng Zhu, Yang Yang and Christopher K. Hsee in the Journal of Consumer Research tests it directly. Participants choose one of two tasks. The tasks are identical in difficulty and length; only the payoff differs. In the urgency condition, the lower-paying task is shown with a short window. For example, one task expires in 10 minutes and the other in 24 hours. But the urgency is spurious: the task can be finished before the deadline in both cases. Participants are also told up front that they can perform only one task. That removes the “I will finish the urgent one first, then move to the important one” rationale.
Table 1. The mere urgency effect: share choosing the lower-paying task (verified data)
| Experiment | Participants | Payoff difference | Urgency condition | Control condition | Gap (percentage points)* |
|---|---|---|---|---|---|
| Experiment 1 (students, lab) | 124 | 5 vs 3 chocolates | 31.3% | 13.3% | 18.0 |
| Experiment 2A (MTurk workers) | 203 | 16 vs 12 cents | 35.3% | 13.9% | 21.4 |
| Experiment 2B (MTurk workers) | 204 | 16 vs 12 cents | 48.1% | 7.3% | 40.8 |
| Experiment 3, no payoff reminder | 402 (total, 4 cells) | 16 vs 12 cents | 33.7% | 9.8% | 23.9 |
| Experiment 3, payoff reminder | 402 (total, 4 cells) | 16 vs 12 cents | 20.6% | 15.0% | 5.6 (not significant) |
| Experiment 4, at least once in three rounds | 402 | Lottery entry for a 25 vs 20 dollar gift card | 58.8% | 38.4% | 20.4 |
Source: Zhu, Yang and Hsee (2018), Journal of Consumer Research 45(3). Percentages are from the paper; the “Gap” column is a CEOtudent calculation (urgency minus control).
Three lessons follow from the table:
- The effect is consistent. In all five comparisons without a payoff reminder, the urgency condition raised the share choosing the lower-paying task by 18 to 41 percentage points. The average gap is 24.9 percentage points (CEOtudent calculation).
- The antidote is simple. In Experiment 3, when each task’s payoff was restated at the moment of choice, the effect was no longer statistically significant. In the urgency condition, the share choosing the lower-paying task fell from 33.7% to 20.6%. According to the authors, the mechanism is attention: a short window grabs attention and pulls it away from the size of the outcome.
- Feeling busy makes it worse. In Experiment 4, the effect was stronger among people who saw themselves as busier in general. The person who most needs the matrix is the one most likely to forget to use it.
A limitation worth noting: the experiments used small payoffs with students and online workers. They do not measure how much value is lost in real working life. But the direction is clear: even a fake deadline can shift choices.
The new owner of the “delegate” box in 2026: AI agents
In 1989, “delegate” meant a colleague. Today the same box can hold an AI agent or a simple automation. The WORKBank study by Shao and colleagues at Stanford University, built on the U.S. Department of Labor’s O*NET database, collected preferences from 1,500 workers and capability assessments from AI experts for 844 tasks across 104 occupations.
Table 2. What do workers want from AI agents? (verified data)
| Finding | Value |
|---|---|
| Share of tasks where workers are positive about automation | 46.1% |
| Most common reason among supporters: “free up time for high-value work” | 69.38% |
| Reason: the task is repetitive or tedious | 46.6% |
| Reason: it would improve work quality | 46.6% |
| Reason: the task is stressful or mentally draining | 25.5% |
| Participants who expressed concerns about using AI | 28.0% |
| Most common concern in that group: lack of trust in accuracy and reliability | 45.0% |
| Share of occupations where “equal partnership” (H3) is the dominant preference | 45.2% |
| Tasks where workers and experts chose the same level of human involvement | 26.9% |
Source: Shao et al., Future of Work with AI Agents, Stanford University, arXiv:2506.06576 (v3, February 2026).
This table speaks directly to the Eisenhower matrix. The number one reason workers want automation is to free time for important but not urgent work, the matrix’s second box. Automation is not the goal in itself. The goal is to win back time for thinking, learning and building relationships.
The study also carries a warning. In 45.2 percent of occupations, the dominant preference is not full automation but an equal partnership between human and agent. And the top concern among those who worried was reliability. So “delegate” is no longer one decision. You have to choose whether the work goes to an agent, to a rules-based automation or to an approved partnership. If you want to go deeper on that boundary, see our guide on whether you should let AI make the decision and our analysis of which knowledge-work tasks to automate first.
The CEOtudent matrix: four boxes, four owners
Leading yourself like a CEO means every piece of work has an owner. Learning like a student means your best time goes to learning and thinking. The table below translates the classic matrix into 2026 conditions.
Table 3. The 2026 Eisenhower matrix (CEOtudent editorial framework)
| Box | Classic action | 2026 owner | Decision rule |
|---|---|---|---|
| Urgent and important | Do it now | Human decides, agent prepares | If the outcome is irreversible or directly affects people, the final call stays with you. The agent prepares summaries, drafts and options; it does not approve. |
| Important, not urgent | Schedule it | Human only (AI can be a study partner) | Set at least one protected calendar block per week and an external deadline. Strategy, learning and relationship work live here. This box is not delegated to an agent. |
| Urgent, not important | Delegate it | Automation or approved agent | If the work is repetitive, rules-based and reversible, automate it. If it needs judgment, let an agent do it and you approve. If neither fits, delegate to a person. |
| Neither urgent nor important | Delete it | Nobody | If nobody notices its absence for 30 days, it stays deleted. Also switch off the notifications and automations that generate this work. |
| Every new “urgent” request | (Not in the classic matrix) | Human, 10-second test | Before deciding, ask two questions: “What do I lose if this is done 24 hours later?” and “What is the payoff?” If the answer is unclear, the work is not important; it only looks urgent. |
The last row rests directly on the Zhu, Yang and Hsee finding: making the payoff visible at the moment of choice was the intervention that weakened spurious urgency. In practice, that means something simple. Next to every item on your to-do list, write the outcome, not just the deadline.
Practice: a weekly 20-minute matrix ritual
- Empty (5 minutes). Write every task in your head and your inbox into one list.
- Write the outcome (5 minutes). Next to each item, note in one sentence what it will gain you or what it will cost you if it is not done. An item with no writable outcome probably belongs in box four.
- Sort into boxes (5 minutes). Apply the rules in Table 3. Put every item labeled “urgent” through the 24-hour test.
- Assign an owner (5 minutes). For each item in box three, choose automation, agent or person. Block time in your calendar for at least one item in box two.
Set yourself an honest measure too: at the end of the week, how much of your time went to box two? If that share is not rising, automation is not saving you time; it is only generating more urgent work. A self-led CEO notices this. A student who never stops learning invests the freed time in a new skill.
FAQ
Did Eisenhower invent the Eisenhower matrix?
No. In his 1954 speech, Eisenhower quotes the urgent versus important distinction from an unnamed former college president, and the speech contains no matrix. The four-box structure was popularized by Stephen R. Covey’s 1989 book.
What is the “mere urgency effect”?
It is the tendency to choose a task simply because its window looks short, even when its payoff is clearly lower. Zhu, Yang and Hsee (2018) demonstrated it in five experiments. The cause is attention: a short window captures attention and pushes the size of the outcome into the background.
What is the single most effective step against urgency bias?
Make the outcome visible at the moment of choice. In Experiment 3, when each task’s payoff was restated, the effect was no longer statistically significant. Writing the outcome next to the deadline on your to-do list is the everyday version of this.
Which tasks should I delegate to an AI agent?
Urgent but unimportant, repetitive and reversible work. In Stanford’s WORKBank study, the top reason workers supported automation was freeing up time for high-value work. For decisions that are irreversible or directly affect people, the agent should prepare and the final call should stay with you.
Why do we keep postponing important but not urgent work?
Because that work has no alarm of its own, and attention drifts to whatever is about to expire. According to Experiment 4, people who see themselves as busier are more exposed to this effect. The fix is to give this work an external deadline and protected time on the calendar.
Sources
- Eisenhower DD. Address at the Second Assembly of the World Council of Churches, Evanston, Illinois, 19 August 1954. The American Presidency Project, University of California, Santa Barbara.
- Zhu M, Yang Y, Hsee CK. The Mere Urgency Effect. Journal of Consumer Research, 45(3), 673-690, 2018.
- Covey SR. The 7 Habits of Highly Effective People. Simon & Schuster, 1989.
- FranklinCovey. Habit 3: Put First Things First. The 7 Habits of Highly Effective People course page.
- Shao Y, Zope H, Jiang Y, Pei J, Nguyen D, Brynjolfsson E, Yang D. Future of Work with AI Agents: Auditing Automation and Augmentation Potential across the U.S. Workforce. Stanford University, arXiv:2506.06576, v3, 2026.
Tables 1 and 2 are verified data taken from primary sources. The “Gap” column in Table 1 and the average gap in the text are CEOtudent calculations. Table 3 is a CEOtudent editorial framework; it is a decision proposal based on this data, not a research finding.
This content was compiled with the support of AI following in-depth research, then written and prepared for publication by the CEOtudent editorial team.
This post is also available in:
