# How to run a Lightning Decision Jam workshop and why AI facilitation helps _A practical guide to moving a team from scattered problems to a realistic, prioritised course of action in 80 minutes._ Some team problems are not short of opinions. They are short of a route from opinions to a decision. The same concerns return in planning meetings, proposed fixes multiply, and the group eventually agrees to continue the discussion elsewhere. This is tactful, but not terribly productive. A Lightning Decision Jam gives that conversation a firm sequence: collect the issues, choose what matters, reframe the chosen problem, then compare possible actions by impact and effort. The 80-minute format below works for 2 to 10 people and produces a decision direction rather than another handsome collection of notes. ## Prepare a decision worth making Give the session a clear boundary. “How do we improve delivery?” is too large. “What should we change to reduce avoidable delays between code review and release?” is specific enough to examine. State what the group can decide, any fixed constraints, and who will own the result. Do not invite people to choose if the answer has already been settled elsewhere. Invite participants who see different parts of the problem. Ask them to bring examples, not prepared speeches. Useful evidence might include customer feedback, incident records, delivery data or a recent case. A Decision Jam can structure judgement, but it cannot make a vague hunch more factual by placing it on a board. ## 1. Gather the field with Round Robin Use 20 minutes for Round Robin. Give each person a turn to contribute one concrete problem, candidate fix or action constraint. Continue around the group until the useful material is visible. Ask short clarifying questions: when did this happen, who was affected, and what makes it costly? Keep one idea per contribution so later voting remains intelligible. > Name something the team could recognise in practice. We need events and effects, not a general sense that the process is a bit wonky. > — A prompt for more useful contributions Turn-taking prevents the fastest speaker from setting the entire agenda. It should not become an interrogation or force everybody to invent a novel point. Passing is allowed. The aim is broad collection, including dissent and constraints, before the group starts defending solutions. ## 2. Narrow the problem with a Blind Vote Spend 10 minutes choosing which contributions best support clear next steps. Set the criterion before collecting votes: importance, confidence in the evidence, and usefulness after the workshop. Let people vote privately, then reveal the result together. Private voting does not remove authority, but it gives the decision owner a less distorted view of the room. Select a manageable number of problems for the next stage, usually one or two. Record close alternatives and material objections. A vote is a prioritisation signal, not proof that the minority is wrong. If a sponsor chooses against the result, note why. Hidden overrides are an efficient way to make the next vote decorative. ## 3. Reframe with How Might We questions Use 25 minutes to turn the selected problems into How Might We questions. Good questions remain anchored in the evidence while allowing several responses. “How might we stop release delays?” is broad. “How might we help reviewers identify release-blocking changes before the final approval?” gives the team a useful target without smuggling in a preferred feature. Write independently first, then compare and combine genuine duplicates. Test each question: does it name the relevant moment or person, permit different answers, and sit within the team’s influence? Choose the strongest frames and generate candidate actions. Keep ambitious ideas visible, but distinguish them from actions the team could actually start. ## 4. Compare actions on an Impact Effort Map Give the final 25 minutes to mapping candidate actions by expected impact and implementation effort. Agree what impact means for this decision and make estimates relative, not theatrical. Discuss differences in placement because they often reveal hidden assumptions about dependencies, risk or who must do the work. High-impact, lower-effort options are sensible starting points, not an automatic answer. A difficult action may address the real cause while an easy one merely tidies the edges. Finish by naming selected actions, an owner, a trigger or start date, and a checkpoint. Park the rest explicitly so nobody leaves believing every sticky note is now a commitment. ## Why AI changes the Decision Jam This workshop looks simple on paper, but its value depends on the handoffs. Somebody must collect several participants’ contributions, protect quieter voices, close each timebox, summarise without erasing disagreement, run a private vote and carry the selected problems into reframing. They must then move candidate actions into the map and make the final commitments usable. ### Reliable facilitation without a routine consultant budget A skilled consultant is worthwhile for a fraught or consequential decision. Few teams have the budget to hire one whenever an ordinary operational problem gets stuck. The usual substitute is a template operated by the product manager, engineering manager or whoever made the mistake of being visibly organised. Clarvelo is the AI facilitator, not an assistant whispering suggestions to a human chair. It introduces the exercises, gathers input from everyone, manages transitions and timeboxes, runs voting, synthesises the material and prompts for actionable outputs. That makes reliable process facilitation practical for routine team decisions, not only special occasions with a consultancy line in the budget. ### The diligent organiser can join the decision The person who organises the workshop often holds important context. When they also explain Round Robin, monitor speaking order, watch four clocks, prepare the vote and rearrange the map, their attention is spent on operating the room. Their contribution arrives late or not at all. With Clarvelo running the process, that organiser contributes and votes on the same terms as everybody else. Simultaneous collection gives reflective and quieter participants room before discussion takes over. Structured synthesis keeps minority concerns attached to the options, while explicit transitions show how one exercise’s result becomes the next exercise’s input. AI facilitation does not supply evidence or make political trade-offs disappear. Serious conflict and high-stakes decisions may need independent human judgement. For the recurring decisions that otherwise drift through three meetings, a consistent facilitator and a visible output are a considerable improvement. ## An 80-minute Lightning Decision Jam agenda - 20 minutes: collect problems, fixes and constraints with Round Robin - 10 minutes: select the most useful issues with a Blind Vote - 25 minutes: reframe selected problems with How Might We questions - 25 minutes: compare actions on an Impact Effort Map and assign follow-up ## Turn the next stuck discussion into a decision Clarvelo’s [Lightning Decision Jam](/workshops/lightning-decision-jam) guides 2 to 10 people through Round Robin, Blind Vote, How Might We and an Impact Effort Map. [Start the AI-facilitated workshop](/workshops/lightning-decision-jam) when your team needs a practical direction and everyone needs a seat in the conversation. --- Written by Joe Corcoran for [Clarvelo](https://clarvelo.com). Clarvelo helps teams run structured retros, decision jams, post-mortems, and other key decision-making workshops with an AI facilitator. Assigned outcomes are automatically exported to wherever your team works. [Read more on the Clarvelo blog](https://clarvelo.com/blog) or [try the live AI-facilitated workshops](https://clarvelo.com).