The creative brief has grown thinner in most companies over the years, and with generative tools that thinness takes its revenge immediately. Someone who briefed vaguely used to get one mediocre idea back. Today they get thirty, all plausible, and none of them decidable. That has changed my view of this single document more than any debate about tools has.
- The faster drafts appear, the more a weak brief costs.
- A brief is a decision, not a wish list.
- Five points belong in every creative brief.
- Generated drafts move the work to the front.
Table of contents
Why the creative brief matters again
For a long time the brief was a distribution document. It said who delivers what by when, and the actual work began afterwards. That held as long as execution was expensive. Anyone who wanted a shoot had to plan it, book it and pay for it, and those costs limited the number of attempts all by themselves. Exactly that limit has fallen away. What has become scarce is no longer production but judgement.
The second reason lies in volume. A team working from a vague instruction now produces more variants in an afternoon than a review can assess. Volume does not create a choice, though, it creates fatigue. By the twentieth version it is no longer the best argument that decides but the clock. A good creative brief is therefore less a set of instructions for starting than an agreement about when to stop.
Then there is a side of this that nobody in marketing was thinking about until recently. The European Commission sets out in its regulatory framework for artificial intelligence which transparency obligations apply to generated content and from when. That is not a legal detail for the legal department but a question that belongs in the creative brief, because it decides how something is made rather than whether it is approved at the end.
The third point concerns authorship. When a draft comes out of a tool, nobody is initially responsible for it being right. That sounds harmless and is not, because responsibility dissolves across rounds in which everyone contributed something and nobody decided anything. A brief that leaves open who signs at the end produces exactly those rounds.
Five points I ask for in every creative brief
The five points below are a sequence, not a form. They come out of projects where the work went well and out of ones where it dissolved into variants.
- Decision. What decision is this work supposed to trigger? Anyone who cannot say that in one sentence has been briefing an occupation.
- Exclusion. What should explicitly not come back? One exclusion saves more time than three examples of what is meant.
- Reference. What is the draft measured against, and what not? Without a reference the loudest taste in the room wins.
- Provenance. What may be generated, what has to be photographed? That question belongs at the start, not in the approval.
- Sign-off. Who decides, and after how many rounds? Without a limit every round is the second to last.
The second point is the one I underestimated longest. An exclusion feels impolite to write because it sounds like mistrust. In collaboration it works the other way round. It takes routes out of the room that nobody wanted to walk anyway and makes space for the ones worth arguing about. Since then every creative brief of mine contains at least two sentences about what is not meant. How I judge the drafts afterwards I have described in evaluating creative work.
What I learned handling generated drafts
Some time ago we changed the first round. It happens without images. What gets presented is only what an idea claims, in a few sentences, and the second round is the first visual one. That was unpopular at first, because a sentence sells worse than a good picture. After a few projects the effect was unambiguous. The weak ideas fell earlier, and the strong ones came through with fewer rebuilds. A creative brief that fixes this order makes more difference than any instruction about image language.
A smaller change came with it that returned surprisingly much. We now record which version we discarded and why, in two lines, straight after the review. It costs a few minutes and prevents the same idea from surfacing three weeks later and being discussed again. In projects with many people involved that was the point at which the number of loops fell noticeably.
The second observation concerns tone. Generated drafts hit the average of a category remarkably well, and that is precisely the problem. They look like what already exists, because they come from it. Anyone who wants to keep a brand distinguishable therefore has to name in the creative brief what it is unlike, and not only what it stands for. Something similar applies to visibility in models, which I described in visibility in language models.
Less of this can be measured than I would like. The quality of an idea remains a judgement. What can be counted is the path towards it: how many rounds until the decision, how many days between instruction and approval, how often an approval was reopened later. Those three numbers say honestly whether the creative brief did its work. They say nothing about the effect of the campaign, and I do not claim they could.
Where the creative brief runs into limits
There are two cases in which a better document saves nothing. The first is an instruction with no internal agreement behind it. A creative brief can write a position down, it cannot manufacture one. Where people disagree, the disagreement comes back in the drafts, only more expensively. The second case is a team without the time to read what is written there at all. A creative brief nobody reads is a safeguard rather than leadership. In both cases the more honest answer is a conversation rather than a longer document, and before anybody starts designing.
The second limit is a matter of craft. Tools change faster than processes get agreed. Anyone writing requirements today that are cut to fit a particular program will have a document to maintain within months. So I consider anything that names tools risky and anything that names intentions stable. A creative brief should still read well when the technology underneath has been swapped out.
In the end the creative brief is not a new discipline but the oldest task of leadership in short form. It says what holds, what does not hold and who decides. Generative tools have changed none of that, they have only raised the price of vagueness. Take the document seriously and you spare yourself the later discussion about why thirty good drafts together produced no decision.