# How to run a Five Whys workshop and why AI facilitation helps _A practical guide to tracing a recurring problem beyond its visible symptom and turning the likely causes into useful opportunities for action._ Five Whys has an appealingly plain premise: state a problem, ask why it happened, then keep asking until the group reaches something deeper than the first explanation. It can uncover a useful cause in half an hour. It can also become one person repeatedly asking ‘why?’ until everyone feels they are being interviewed by a determined four-year-old. The difference is in the facilitation. A good Five Whys workshop follows evidence, distinguishes causes from restated symptoms, and treats the resulting chain as a plausible explanation to test rather than revealed truth. This 60-minute format pairs Five Whys with How Might We, so the group moves from investigation to opportunities for improvement. ## Before the workshop Choose one concrete, recurring problem. ‘Customers are unhappy’ is too broad. ‘A third of new workspace owners abandon setup before inviting a colleague’ gives the group an observable event to examine. Bring people who saw different parts of the problem, along with incident notes, support examples, product data or process records. Memory is useful, but it improves considerably when introduced to evidence. - Describe what happened without building a preferred cause into the wording. - Set a boundary, such as one incident, process or customer journey. - Invite people close to the work, not only people accountable for it. - Agree what follow-up decision or experiment the session should inform. ## 1. Establish the problem and a blameless tone Open by reading the problem statement and checking that participants recognise it. Clarify the effect, frequency and relevant conditions. Explain that the exercise is investigating how the system produced an outcome, not searching for the person whose name can be placed at the bottom of a diagram. If an answer is ‘someone made a mistake’, ask what made that mistake possible or likely. > We are building a cause chain we can test, not conducting a trial. Each answer should explain the one before it and stay close to what we know. > — A sensible opening agreement ## 2. Ask the first why independently Give participants quiet time to answer why the stated problem occurs. Collect responses from everyone before discussion. This prevents the first confident explanation from setting the entire chain and gives quieter people, who may hold the awkward operational detail, an equal route into the analysis. Ask for a specific example or piece of evidence where an answer is vague. Several answers may reveal several causal branches. Do not force them into one neat sentence merely because the method has a singular name. Choose the branch most relevant to the workshop’s scope, record alternatives, and be explicit about why the group is following this one. ## 3. Build and check the cause chain For each successive round, ask why the previous answer is true. Compare contributions, synthesise the common explanation without erasing meaningful disagreement, and check the link before continuing. ‘Setup instructions are unclear’ followed by ‘customers are confused’ has travelled in a circle. A stronger next answer identifies what makes the instructions unclear, perhaps role-neutral wording or missing examples. 1. Does this answer explain the previous one? 2. What observation or evidence supports it? 3. Are we describing a system condition rather than assigning blame? 4. Could another cause also produce the same effect? 5. Is this cause within the team’s influence? Five is a prompt, not a law of nature. Stop at three if the group reaches a specific, useful cause. Continue beyond five if the chain remains superficial and time allows. Stop altogether when evidence runs out, mark the uncertainty, and name what must be investigated. A workshop should not reward persistence with invented certainty. ## 4. Turn causes into How Might We questions Use the second 30 minutes to reframe the useful causes as opportunities. Ask each participant to write How Might We questions independently. ‘How might we help workspace owners recognise the setup steps relevant to their role?’ opens several possible responses. ‘How might we add a checklist?’ has smuggled a solution into the question before alternatives have had their coats off. Group genuinely similar questions, then discuss which are specific enough to guide work and broad enough to permit different solutions. Finish with a concise set of opportunity questions and state where they go next: discovery, an experiment, prioritisation or delivery planning. Keep unselected causal branches and uncertainties in the summary so useful dissent does not vanish. ## Why AI changes this workshop Five Whys looks simple on paper, yet somebody must collect several participants’ answers at every round, compare them, test the causal link, protect the timebox and make a clean transition into How Might We. A static template cannot prevent circular reasoning or carry a qualified cause into the next activity. ### Reliable process without a routine consultant budget A skilled external facilitator is valuable for a serious incident, damaged trust or a politically fraught diagnosis. Most teams cannot hire one whenever a recurring operational or product problem needs an hour of disciplined attention. Clarvelo makes reliable process facilitation available for those ordinary, important sessions. It is the AI facilitator: it runs the input rounds, asks the next question, keeps time and guides the group between exercises. ### Everyone gets to investigate Without active facilitation, the diligent organiser becomes chair, timekeeper, note-taker and cause-chain custodian. That person is often closest to the problem and should be contributing evidence, not preparing the next prompt while colleagues think. When Clarvelo runs the room, the organiser participates on the same terms as everyone else. Multi-participant collection matters here. Clarvelo gathers each person’s view before discussion, making quieter voices visible and reducing anchoring. It synthesises points while preserving alternatives, checks transitions between why rounds, then carries the resulting causes into How Might We. The final output records the cause chain, evidence, uncertainty and opportunity questions, rather than leaving behind an attractive row of boxes with no agreed use. AI does not make the diagnosis true. The team remains responsible for evidence and judgement, and some problems need specialist investigation. Clarvelo provides the discipline that helps a group know what it found, what remains uncertain and what should happen next. ## A practical 60-minute agenda - 30 minutes: Five Whys, with separate input rounds and cause checks - 30 minutes: How Might We, from causal insight to useful opportunity questions ## Investigate the cause without operating the workshop Clarvelo’s [Five Whys Deep Dive](/workshops/five-whys-deep-dive) guides 1 to 8 people through Five Whys and How Might We in one connected 60-minute session. Bring one concrete problem and the people who understand it; Clarvelo will run the process and produce a usable follow-up frame. --- Written by Joe Corcoran for [Clarvelo](https://clarvelo.com). Clarvelo is an AI facilitator that guides teams through retrospectives, decision making, problem solving, prioritization and 40+ workshop formats. [Read more on the Clarvelo blog](https://clarvelo.com/blog) or [try the live AI-facilitated workshops](https://clarvelo.com).