From brief to boards in an afternoon, without guessing what they meant

When visual boards required either illustration or a long search through stock, they were produced once, late, and defended. The cost of making them meant they arrived as a proposal rather than as a tool for thinking.

When a board takes an afternoon, it becomes something else: a way of asking a question. You can produce three directions instead of one, show them early, and let the client point rather than describe.

The risk is producing a great deal of attractive material that answers the wrong question. Speed only helps if the extraction step at the front is done properly.

Go through the brief for hard constraints, stated negatives, and any reference with a note on what specifically is being referenced. This is the same extraction any brief deserves and it matters more here because production is about to become fast.

Fast production without a constraint list produces a large volume of confidently wrong work. The client sees ten frames that miss the point and concludes you did not read the brief — which, functionally, you did not.

Write the list. Keep it visible while you work. Check candidates against it before they go on a board.

Produce three genuinely distinct territories. They should differ in something structural — the setting, the treatment, the level of realism — so that choosing between them is a real decision.

Three near-identical options force the client to articulate a preference they do not have, and the feedback comes back as vague adjustments. Three real alternatives get a clear answer, and often a useful hybrid: 'the second one, but with the light from the third'.

Label them plainly and neutrally. Naming one 'the safe option' tells them what you think and biases the answer you get.

Each direction needs enough frames to establish that it is a coherent world and not a lucky image. Six to nine does that. Twenty does not do it better; it dilutes, and it invites feedback on individual frames rather than on the direction.

Include a mix: an establishing frame, two or three that carry the treatment, one detail, and one that shows the hardest shot in the brief. That last one is important — boards that avoid the difficult shot are pleasant and dishonest.

Lay them out on one page per direction so the whole thing is seen at once. A board viewed as a sequence of full-screen images reads as a series of separate pictures, which is exactly what you are trying to avoid.

Generated boards can promise shots that will be difficult or impossible to deliver — a complex interaction, readable type, a specific real location. Approving those creates an expectation you then have to meet.

Mark the frames that are indicative rather than achievable, and say what the production version would involve. It is a small note and it prevents the worst conversation in the job, which is the one where the finished work is compared unfavourably to the board.

Clients respond well to this. Someone who says 'this frame is the intent, the real version would be shot differently' reads as experienced, not as hedging.

Keep the prompt beside every board frame. When a direction is approved, those prompts are the starting point for production — the treatment is already locked, the craft block is already tested.

This is the real advantage of generated boards over sourced ones. A stock-based board tells you what to aim for; a generated board hands you the mechanism that produced it.

Structure the file so this is automatic: frame image, prompt, and the note about what is achievable, together. Approval then converts directly into a shot list rather than requiring a translation step.