Critical thinking questions: 100 for work, plus 20 exercises
Sorted by what you need them for, from clarifying a problem to reviewing a project. Plus team exercises, examples and interview questions.
Critical thinking at work means checking an idea before you act on it: what the problem really is, what you are assuming, how good the evidence is and what could go wrong. The quickest way to do that is to ask better questions. Below are 100 critical thinking questions for work, grouped by the moment you need them, then 20 exercises that build the habit in a team, 8 short examples of critical thinking at work and 10 interview questions that test it.
100 critical thinking questions for work
Pick the group that matches where your team is: defining a problem, testing a plan, choosing, or looking back. Two or three questions are enough for one meeting.
Questions that clarify the problem
Ask these first. Most bad decisions answer the wrong question well.
1. What exactly is the problem we are trying to solve?
2. How would we describe this problem to someone outside the team in one sentence?
3. Who is affected by this, and how do we know?
4. When did we first notice it, and what changed around that time?
5. Is this the problem, or a symptom of a bigger one?
6. What would be different if this problem were solved?
7. What is in scope here, and what are we deliberately leaving out?
8. Whose problem is this to solve?
9. How big is this problem compared with everything else on our plate?
10. Which words in the problem statement could two people read differently?
11. Is the deadline real, or did we set it ourselves?
12. What have we already tried, and what happened?
13. What would happen if we did nothing at all?
14. Which single question, once answered, would make the rest easy?
15. Are we solving the problem we were asked to solve, or the one we find more interesting?
Questions that test assumptions
Every plan rests on things nobody has checked. These bring them into the open.
16. What are we taking for granted here?
17. What would have to be true for this plan to work?
18. Which of those things have we actually checked?
19. Where did this belief come from: data, experience or something someone said once?
20. Would this still hold with twice as many customers, or half the budget?
21. What if the opposite were true?
22. Are we assuming this because it is likely, or because it is convenient?
23. Which assumption, if it turned out wrong, would hurt us most?
24. How could we test that assumption cheaply this week?
25. Are we treating a one-time event as a pattern?
26. What do we assume our customers want, and when did we last ask them?
27. Are we keeping this step because it works, or because it has always been there?
28. What are we assuming about how long this will take?
29. Are we assuming people have the time and skills this needs?
30. Which rule are we following that nobody remembers making?
Questions that weigh the evidence
Not every number or story deserves the same weight. These sort strong evidence from weak.
31. How do we know that?
32. What is the source, and how close is it to the original data?
33. How many examples is this based on?
34. Compared with what: last month, last year or another team?
35. Is this a fact, an opinion or a forecast?
36. What evidence would change our minds?
37. What evidence have we not looked for yet?
38. Could something else explain the same result?
39. Did one thing cause the other, or did they just happen at the same time?
40. Who is missing from this data?
41. Is this an average that hides a wide range?
42. Who collected this, and what did they hope it would show?
43. How old is this information, and has anything changed since?
44. Would we accept this evidence if it pointed the other way?
45. What is the strongest evidence against our view?
Questions that bring in other perspectives
A room full of people who agree is a warning sign. These borrow the view of someone who is not in it.
46. How would a brand-new customer see this?
47. How would the person who has to carry this out describe it?
48. Who loses something if we do this?
49. What would our most skeptical colleague say?
50. How would someone in finance, legal or support see this?
51. Who hasn't spoken yet, and what do they think?
52. What is the best argument for the option we are about to reject?
53. How will this look to someone who joins the team next year?
54. What does the other team know that we don't?
55. If we had to explain this to the people it affects, which part would be hardest to say?
56. Where could our own experience be misleading us here?
57. What would someone from a completely different industry do?
58. Are we hearing from the people who use this every day, or only from the people who manage it?
59. Which voices are loudest in this discussion, and do they speak for most people?
Questions that think through consequences
Good ideas still have side effects. These look past the first step.
60. If this works, what happens next?
61. And what happens after that?
62. What is the worst realistic outcome, and could we live with it?
63. Who will feel the effect first?
64. What will this make harder later?
65. What will we have to stop doing to make room for this?
66. How will this look in six months?
67. Can we undo this if it goes wrong, and at what cost?
68. How might people work around it?
69. What does it cost us to be wrong in each direction?
70. Which side effect are we hoping nobody notices?
71. What does this commit us to beyond this project?
72. How will this affect the people who are not in this room?
73. What new problem could this solution create?
Questions for making the decision
For the moment the team has to choose. They slow the room down just enough.
74. What are we actually deciding today?
75. What are our options, including doing nothing?
76. Which criteria matter most, in order?
77. Did we agree on the criteria before we looked at the options?
78. What would make us more confident, and is it worth waiting for?
79. Which option would we pick if we had half the time?
80. Who makes the final call, and who needs to be consulted first?
81. How confident are we on a scale of 1 to 10, and why not higher?
82. What would make us revisit this decision?
83. Are we choosing this because it is best, or because we have already spent so much on it?
84. Is there a small version we could try first?
85. What would we advise another team to do in our position?
86. What is the simplest option that would work?
87. Will we still be able to explain why we chose this a year from now?
Questions for an after-action review
Ask these once the project ships or the dust settles, while people still remember the details.
88. What did we expect to happen?
89. What actually happened?
90. Why was there a difference?
91. Which parts went well because of our plan, and which because of luck?
92. What did we know at the time, and what do we only know now?
93. When was the first sign that something was off, and who saw it?
94. Which of our assumptions turned out to be wrong?
95. What would we do the same way next time?
96. What would we do differently?
97. What should we write down so the next team does not have to learn this again?
98. Did we decide at the right moment, or too early or too late?
99. Who else should hear what we learned?
100. What is one change we will make starting this week?
How to ask without sounding like an interrogation
The same question can feel like curiosity or like an attack. The difference is mostly in how you set it up.
- Say why you are asking. "I want to stress-test this before we commit" turns a question into help. Without it, people hear doubt.
- Ask about the idea, not the person. "What would have to be true for this to work?" invites thinking. "Why do you think that?" can sound like an accusation.
- Lead with what and how. "What are we assuming?" and "How would we know?" open a conversation. Save why for when the room is relaxed.
- Ask one question, then wait. Count to five in your head. A stack of three questions in a row feels like a cross-examination, and people answer the easiest one.
- Question your own idea first. When the most senior person says "here is where my plan is weakest", everyone else feels free to do the same.
- Make it a role, not a personality. Put two or three questions from this page in the agenda, or rotate who asks them each meeting, so nobody becomes the official skeptic.
Where these questions come from
Socratic questioning is a way of examining an idea through a series of open questions instead of arguing against it. It is named after the Greek philosopher Socrates, who questions people this way in Plato’s dialogues. A common teaching version, usually credited to the critical thinking educator Richard Paul, sorts the questions into six kinds: questions that clarify, questions that probe assumptions, questions that probe reasons and evidence, questions about other viewpoints, questions about implications and consequences and questions about the question itself. The groups above follow the same idea, adapted to a workplace.
20 critical thinking exercises for teams
These critical thinking activities each list the time, the group size, the steps and two debrief questions. Spend at least a third of the time on the debrief: that is where the habit forms.
Quick exercises for a meeting
Ten minutes or less, and all of them work on a video call.
1. Out of how many?
Builds: Asking for the base before reacting to a number.
10 minutes · Any size · Five statements on a slide · works on video
- Write five statements that each carry a number with no base, such as "Complaints doubled this month" or "Three customers asked for a dark mode."
- For each one, the group says what it needs to know before deciding whether it matters: out of how many, compared with when, measured how.
- Reveal the context you prepared for each. For example, complaints rose from 2 to 4 out of 3,000 orders.
Debrief
- Which statement sounded most alarming before you heard the context?
- Where do we share numbers without a base in our own reports?
2. Three-finger sort
Builds: Telling what is known from what is believed or expected.
10 minutes · Any size · A status update with twelve sentences pulled out · works on video
- Pull twelve sentences from a recent status update, such as "Sign-ups rose 8% last week", "The new design is cleaner" and "We will finish by June."
- Read them one at a time. Everyone holds up one finger for fact, two for opinion, three for prediction. On a call, type F, O or P in the chat.
- Talk only about the sentences where the room split.
Debrief
- Which kind of sentence was hardest to classify?
- How much of our last real update was opinion written to sound like fact?
3. What would change your mind?
Builds: Holding views that evidence could actually move.
10 minutes · Pairs · No materials · works on video
- Each person picks a low-stakes opinion they hold about work, such as "Morning meetings are better than afternoon ones."
- They write down the specific evidence that would make them drop it.
- Partners swap notes and judge whether that evidence could really be found. If it couldn't, rewrite the opinion until it could be tested.
Debrief
- Was it harder to name the evidence for the opinions you hold most strongly?
- Which belief on our team could no evidence ever change?
4. Who's missing from the data?
Builds: Spotting conclusions drawn only from the people who made it through.
10 minutes · Any size · No materials · works on video
- Read a claim built on a filtered group: "Every customer we interviewed loves the new checkout." Then add that the interviews were with customers who had just finished checking out.
- Ask who never made it into this data, and what they might say.
- Repeat with two more: "The people who answered our survey like the new office" and "Our longest-serving employees all say onboarding here was great."
- Name the pattern: survivorship bias, or more broadly selection bias.
Debrief
- Which missing group would change the conclusion most?
- Whose voice is missing from the data we use to decide things?
5. Observe before you infer
Builds: Separating what you see from what you conclude.
10 minutes · Groups of 3 to 6 · One busy photo, such as a market or a crowded desk · works on video
- Put one busy photo on the screen.
- For two minutes, everyone writes only what they can see: "a cup with a lid", not "someone is in a hurry."
- Then two minutes on what they infer from what they saw, and one minute on what they wonder about.
- Compare lists and flag every "observation" that was really an inference.
Debrief
- How many of your first observations were already conclusions?
- Where at work do we report an inference as if we saw it?
Exercises for weighing evidence
Fifteen to twenty minutes each. Two come with answers under a fold.
6. Claim, evidence, confidence
Builds: Rating how well a belief is supported before acting on it.
15 minutes · 3 to 10 people · A whiteboard · works on video
- Pick a claim your team repeats often, such as "Customers want faster onboarding."
- Write the claim at the top of the board and list every piece of evidence anyone can name, with its source.
- Everyone rates their confidence in the claim from 1 to 5 in silence, then all reveal at once.
- Agree on one thing you could check this month that would move the score the most.
Debrief
- How far apart were the scores, and what explained the gap?
- Which of our plans rest on a claim nobody would score above 3?
7. Cause or coincidence
Builds: Finding other explanations when two things move together.
15 minutes · Groups of 3 to 5 · One card per group · works on video
- Give each group a card describing two things that move together at a made-up company: "Teams that use the new dashboard close more deals", "Customers who call support cancel more often" or "Projects with a kickoff meeting finish on time more often."
- Groups list at least three explanations other than "A causes B": B causes A, a third thing drives both, the groups were different to begin with, or chance.
- Each group picks its most likely explanation and says how it would test it.
Debrief
- Which explanation surprised you most once you said it out loud?
- What link between two numbers have we treated as cause and effect at work?
8. The too-good pitch
Builds: Spotting weak reasoning dressed up as proof.
15 minutes · Groups of 3 to 5 · The pitch below, printed or on screen · works on video
- Show this made-up sales pitch: "Teams using Tool X are up to 40% more productive. Just read the reviews on our homepage. One support team halved its response time after switching. Every customer we surveyed would recommend us. Everyone is buying productivity tools this year, and our price has never been lower."
- Groups get eight minutes to find six problems with the reasoning.
- Each group reads one problem in turn until all six are found.
Show the answer
1. "Up to 40%" is a best case, not a typical result, and it never says how productivity was measured. 2. Reviews on the seller's own homepage are hand-picked. 3. One team's result is a single story with nothing to compare it to. 4. A survey of current customers leaves out everyone who already left. 5. Lots of people buying a type of tool says nothing about this tool. 6. "Never been lower" compares the price with itself and says nothing about total cost or value.
Debrief
- Which problem was the hardest to spot, and why did it sound convincing?
- Which of these moves do we make in our own pitches and proposals?
9. Two headlines, one table
Builds: Seeing how framing changes what true numbers seem to say.
20 minutes · Two groups of 2 to 6 · No materials · works on video
- Give both groups the same table: 40 people answered a survey about the best day for the team meeting. 24 chose Tuesday, 10 chose Thursday and 6 had no preference.
- Group A writes a headline that makes Tuesday look like a clear winner, such as "60% want Tuesday." Group B writes one that makes it look shaky, such as "16 of 40 did not pick Tuesday." Both must be true.
- Read the headlines aloud, then write one neutral headline together.
Debrief
- Which true headline was more persuasive, and why?
- When we report numbers, where do we choose the framing without noticing?
10. 90% sure
Builds: Learning how wide your uncertainty really is.
15 minutes · Any size · Paper and a pen each · works on video
- Read eight number questions aloud: In what year did the Hubble Space Telescope launch? In what year was the Eiffel Tower completed? How many seconds are there in a day? In what year did the Golden Gate Bridge open? How many minutes are there in a week? In what year did the first iPhone go on sale? How many squares are on a chessboard? In what year did the Panama Canal open?
- For each one, everyone writes a low and a high number so that they are 90% sure the true answer falls between them.
- Reveal the answers. Each person counts how many of their ranges caught the answer. Honest 90% ranges should catch about seven of eight.
Show the answer
In what year did the Hubble Space Telescope launch? 1990. In what year was the Eiffel Tower completed? 1889. How many seconds are there in a day? 86,400. In what year did the Golden Gate Bridge open? 1937. How many minutes are there in a week? 10,080. In what year did the first iPhone go on sale? 2007. How many squares are on a chessboard? 64. In what year did the Panama Canal open? 1914.
Debrief
- If you caught fewer than seven, what made your ranges too narrow?
- Which estimates at work do we give as one number when a range would be more honest?
Prefer puzzles with one right answer? Our problem solving activities include logic grids and ranking challenges, and our escape room puzzle ideas come with solutions.
Exercises for testing assumptions and consequences
Run these on a real plan when you can. They work best when something is at stake.
11. Trace it back
Builds: Finding out where a team belief came from.
20 minutes · 4 to 10 people · A whiteboard · works on video
- Ask for five "facts everyone knows" about your customers or your work, such as "Clients hate phone calls."
- For each one, trace where it came from: who said it first, when, and based on what.
- Mark each one as checked, partly checked or folklore.
- Pick one folklore belief and give someone the job of checking it before the next meeting.
Debrief
- How many of our facts turned out to be one person's story from years ago?
- Which folklore belief is shaping a decision right now?
12. Assumption audit
Builds: Turning a plan's hidden bets into tests.
25 minutes · A project team of 3 to 8 · The current plan for a real project and highlighters · works on video
- Share the current plan for a real project.
- Highlight every sentence that depends on something not yet confirmed: a date, a budget, someone's availability, a customer's reaction.
- Rate each highlighted assumption on two scales: how likely it is to be wrong and how much it would hurt if it were.
- The assumption that scores highest on both gets a test this week, with an owner.
Debrief
- Which assumption had nobody noticed was an assumption?
- What is the cheapest way to test the riskiest one?
13. Weakest link
Builds: Checking each step of an argument, not only its conclusion.
15 minutes · Groups of 3 to 5 · No materials · works on video
- Write the reasoning behind a proposal as a chain of short steps: "Customers are leaving. They leave because setup is slow. A video will make setup faster. They will watch the video. So fewer will leave."
- Each group circles the step it trusts least and explains why.
- Groups then write what evidence would make that step stronger, or what would replace it.
Debrief
- Did every group pick the same weak step?
- Which link in our current plan have we never questioned?
14. And then what?
Builds: Looking past the first effect of a decision.
15 minutes · 4 to 12 people · A whiteboard · works on video
- Write a proposed change at the top of the board, such as "Replace the weekly meeting with a written update."
- Pick three groups it affects. For each, ask "and then what?" three times, writing every answer under the last one.
- Circle each consequence nobody had mentioned before the exercise.
Debrief
- Which circled consequence would change the decision?
- Which of our past decisions would have looked different after three rounds of this?
15. Question storm
Builds: Exploring a problem before anyone proposes a fix.
20 minutes · 3 to 8 people · Sticky notes or a shared doc · works on video
- State a real problem in one sentence.
- For eight minutes, the group writes only questions about it, one per note. No answers, no fixes and no debating.
- Sort the questions, then vote on the three that would most change what you do if you knew the answer.
- Assign each of the three to someone to answer before the next meeting.
Debrief
- How did the problem look different after eight minutes of only questions?
- Which question surprised the person who owns the problem?
Exercises for better decisions
Habits a team can keep after the session, not only games.
16. Criteria first
Builds: Deciding what matters before falling for an option.
20 minutes · 3 to 8 people · A whiteboard or a shared sheet · works on video
- Before anyone sees the options for a real choice, such as a tool, a venue or a vendor, list what matters and weight each criterion out of 10.
- Only then reveal the options and score each one against the criteria.
- Compare the winner with the option people liked best before scoring.
Debrief
- Did the scored winner match the early favorite, and why or why not?
- Which decision coming up would benefit from criteria agreed in advance?
17. The outside view
Builds: Estimating from track records instead of hopes.
20 minutes · 3 to 8 people · No materials · works on video
- Before estimating how long a new project will take, list five past projects that were similar.
- Write down how long each one actually took compared with its plan.
- Make the new estimate starting from that track record, then adjust only for what is genuinely different. Planners call this reference class forecasting.
Debrief
- How far off were our past plans, and in which direction?
- What makes us believe this project will be different?
18. Kill criteria
Builds: Agreeing on when to stop before you are invested.
15 minutes · A project team · No materials · works on video
- At the start of a project, agree on two or three signals that would mean stop or change course, such as "fewer than 20 sign-ups after four weeks."
- Write down who checks each signal and when.
- Put the check in the calendar now, while nobody has a reason to argue for continuing.
Debrief
- Was it easier to agree on stopping points before starting than it would be later?
- Which current project would have hit a kill signal by now?
19. Decision journal
Builds: Judging decisions by their reasoning, not only their outcome.
15 minutes, plus 10 a month later · Any size · A shared doc · works on video
- Before a real decision, each person writes what they expect to happen, how confident they are and their two main reasons.
- Save the notes in a doc nobody edits afterward.
- A month later, compare what happened with the predictions, and then check the reasons separately: a good result can come from bad reasoning.
Debrief
- Were we right for the reasons we gave?
- Whose predictions were best calibrated, and what did they do differently?
20. Add a third option
Builds: Widening a yes-or-no choice before voting on it.
10 minutes · 3 to 10 people · No materials · works on video
- Take a yes-or-no decision on the table, such as "Should we hire a contractor?"
- Before anyone votes, the group lists at least two more options, including doing nothing and doing a smaller version.
- Choose among all of them.
Debrief
- Did the final choice come from the original two options?
- Which yes-or-no question on our list is really a choice between several options?
For decision drills aimed at managers, such as deciding with half the facts, see our leadership activities. For classics like the spaghetti tower and shipwreck ranking, see our team building exercises.
8 examples of critical thinking at work
Short, everyday situations. In each one, a single question changed the decision.
1. The support spike
Support tickets jump sharply the week after a release, and the first instinct is to roll it back.
The question: “What else changed that week?”
What happened: A support lead finds that a new help button made filing a ticket two clicks easier. The release stays, and the team adds answers to the help page instead.
2. Sign-ups doubled
A campaign report says sign-ups doubled, and the team wants to triple the budget.
The question: “Doubled from what, and compared with when?”
What happened: Sign-ups went from 12 to 24, and the week it was compared with was a holiday week. The team runs the campaign for another month before scaling it.
3. The confident candidate
A hiring panel is ready to make an offer to the candidate who interviewed with the most confidence.
The question: “How does each candidate score on the criteria we wrote before the interviews?”
What happened: The favorite scores high on presence and low on the two skills the role needs most. The panel asks the top two candidates for a short work sample before deciding.
4. The cheapest quote
Three software quotes come in, and one is clearly the cheapest for the first year.
The question: “What does each option cost over three years, all in?”
What happened: Once setup, training and the price after the first-year discount are added, the cheapest quote is the most expensive. The team picks another vendor.
5. What customers want
A survey shows customers want a new reporting feature, and it jumps to the top of the roadmap.
The question: “Who did we not survey?”
What happened: Only current customers got the survey. A product manager calls five prospects who chose a competitor, and they describe a different gap, which joins the roadmap discussion.
6. The late deliveries
Deliveries keep arriving late, and one driver is getting the blame.
The question: “Do the late deliveries follow the driver, or something else?”
What happened: An operations manager maps every late delivery and finds they cluster on one route during school pickup hours, whoever is driving. The schedule changes and the driver is cleared.
7. The sure deal
A sales rep calls a big deal 90% certain to close this quarter.
The question: “What has the customer done, not said?”
What happened: There is no signed proposal, no meeting with the budget holder and no date. The forecast is adjusted, and the rep books the meeting that was missing.
8. The lucky launch
A product launch beats its targets, and the team wants to repeat the plan exactly.
The question: “Which results came from the plan, and which from luck?”
What happened: A competitor had an outage that same week. The team keeps the parts of the plan that worked and sets the next target without counting on another outage.
The same habit helps when a team is going through a reorganization or a new system. Our change management activities put it to work.
10 interview questions about critical thinking
Ask for real stories, not hypotheticals, and follow up with “how did you know?” Under each question is what a strong answer shows.
1. Tell me about a time you changed your mind about something at work. What changed it?
A good answer: They name new evidence, not just pressure from someone senior, and they can say what they believed before.
2. Describe a decision you made with incomplete information. How did you decide you knew enough?
A good answer: They separate what they knew from what they didn't, weigh the cost of waiting and explain how they limited the risk.
3. Tell me about a time the data pointed one way and your instinct pointed another. What did you do?
A good answer: They looked for why the two disagreed instead of picking a side out of habit.
4. How would you work out why sales in one region dropped last month?
A good answer: They ask clarifying questions first, list several possible causes and say which data would test each one before settling on an answer.
5. What is a common belief in your field that you think is wrong?
A good answer: A specific position with reasons behind it, plus what would prove them wrong. Contrarian for its own sake is a weak answer.
6. Tell me about a mistake you made because you didn't question something. What do you do differently now?
A good answer: They own the mistake and describe a concrete habit or question they now use, not a vague promise to be more careful.
7. How do you check whether a source you are relying on can be trusted?
A good answer: Concrete checks: who produced it, how, how recently, and whether anything independent confirms it.
8. Tell me about a time you disagreed with a plan your team supported. How did you raise it?
A good answer: They raised it early, with evidence, in the right setting, and then either changed the plan or committed to it.
9. If you had two weeks to evaluate a new tool for the team, how would you go about it?
A good answer: They define what success looks like before testing, try it with real users on real tasks and count costs beyond the price.
10. Tell me about a problem where the first explanation turned out to be wrong. How did you find out?
A good answer: They noticed a fact that didn't fit and followed it, rather than forcing it into the first story.
Want a host to run a thinking game for your team?
An escape room is critical thinking under a clock: clues to weigh, assumptions to drop and a team that has to agree fast. Book a virtual escape room and a live host runs it on Zoom by default, with small teams solving puzzles in about an hour. Book about 2 weeks ahead, or get a free quote.
Four hosted games built on reasoning together, with today’s prices.
