JM Family Design System
Stage 3 · Explore

Ideate

Generate many options, defer judgment, then converge on purpose.

Ideation is where the team gets to explore on purpose. Quality gets easier after the team has enough options. A hundred rough ideas produce better instincts than ten polished ones, because they map the option space.

The discipline is to defer judgment long enough to find the option you would not have thought of in the first ten minutes. When a team converges too early, the workflow collapses into “the first idea everyone could live with.” That is usually a safe compromise, not the strongest direction.

What you do

  • Go wide before you go deep

    Sketch ten directions before evaluating any of them. The first three will be obvious. The next seven are where the work is.

  • Use constraints to provoke, not to censor

    "How would we solve this with only paper?" is generative. "How do we keep this on brand?" is useful. Pick the version of the constraint that opens better ideas.

  • Show, do not tell

    A sketch on paper or a screenshot in chat beats a paragraph in a brief every time. Build the habit of drawing the idea, however ugly.

  • Cluster, vote, then choose

    End the session by collapsing the options into a few directions and picking the ones to prototype. Leave with a decision the team can act on.

What you produce

  • Sketches and concept notes

    Visual artifacts of every direction you considered, including the ones you discarded.

  • Two or three directions worth prototyping

    The shortlist that survived cluster and vote.

  • A written record of what you set aside and why

    The set-aside list saves the next team from re-running this meeting.

AI is a tireless divergent thinker.

Generative AI is unusually good at quantity, which is exactly the part of ideation humans are worst at. Use it to draft fifty variants of an interaction, explore unfamiliar adjacencies, and push back on the obvious move with an unobvious one.

Judgment still belongs to the team. Your job is to look at the wide field AI helped generate and pick the directions that respect what you learned in Empathize and Define. That part stays human.

Watch for

Patterns that can make the work look farther along than it is.

  • Premature convergence

    Falling for the first idea because the meeting feels long. Give the room another twenty minutes of rough options.

  • Critique disguised as evaluation

    "That is not on brand" is critique. "This solves for X but loses Y" is evaluation. Wait for evaluation.

  • Ideating without the brief

    If the team cannot recite the problem statement from Define, the session will drift into solutions for a different problem.

  • Picking by who spoke last

    Use a structured vote, not memory from the last comment. The strongest idea may belong to the quietest person in the room.

What you carry into the next stage

A short list of directions worth making real, plus a written record of the ones you set aside, so the team can keep moving with confidence.

Related in the system