How to Write a Creative Brief in the AI Age

Why the periodic table of creative briefs matters more than ever — and how to turn an old discipline into the strategic core of a modern AI briefing system.

I have wanted to write this essay for years, but I think I finally understand the reason for the timing.

About a decade ago I saw a Picasso retrospective, and it gave me an insight I have repeated to people ever since. Like most reasonably educated people, I thought I knew Picasso. I knew Cubism. I knew the great paintings. I felt I understood the breakthrough. What I had never really considered was what came before the breakthrough.

The exhibit was arranged chronologically, beginning with Picasso’s student work and continuing all the way to the final construction-paper guitars. What struck me immediately was the astonishing realism of the early pieces. There were master studies, beautifully observed heads, and works so technically disciplined that they almost felt like a rebuke to lazy ideas about genius. Before Picasso became something new, he first mastered the old principles at a very deep level.

That lesson has stayed with me because it applies to almost every serious profession. Chemists know the periodic table. Lawyers know the foundations of law. Doctors know anatomy, diagnosis, and physiology before they begin extending the frontier. The new sits on top of the old. Innovation does not abolish fundamentals. Innovation assumes them.

Marketers, by contrast, too often behave as though our discipline has no periodic table at all. We chase fashion. We confuse novelty with mastery. We thrash from framework to framework like headless chickens and call the result agility. That weakness is becoming much more dangerous in the age of AI, because now all those headless chickens are briefing the machines.

The result is not a creative renaissance.

The result is slop at scale.

So this essay is about the periodic table of creative briefs.

By periodic table, I mean the irreducible elements that appear in every useful brief: customer, insight, benefit, proof, character, constraint, and approval. Those elements do not disappear because a new channel arrives, a new agency model emerges, or a new technology starts answering our questions in complete sentences. If anything, they become more important.

A large language model handed a confused audience, a mushy benefit, a hidden objective, twelve competing mandatories, and a fog bank of stakeholder language will do what machines now do very well. It will synthesize the confusion and hand it back in polished form.

That polish is dangerous because it can look like progress.

It is not progress. It is fluent mush.

Every great ad you have ever admired, every campaign you tore from a magazine, sent to a friend, quoted in a meeting, or wished you had made, began with a great creative brief. Not every great brief produces great creative. Plenty of excellent briefs have been betrayed by weak execution, timid clients, nervous committees, or agencies that simply did not find the magic. But a lousy brief has no right to expect rescue.

A bad brief can sometimes be overcome.

It should never be forgiven.

With humility and purpose, then, here is my attempt to write down the periodic table. First I want to show the enduring elements of a great creative brief. Then I want to show how those same elements become the strategic center of a modern AI briefing system. If we learn how to brief properly again, we may be on the edge of a creative renaissance of untold proportions.

So buckle up, students.

You are going to grad school.

---

The short version: a great creative brief answers six questions

Before we get into the history, the lineage, the workshops, the old P&G ghosts, the Green Book, the agency behaviors, the AI stack, the machines, the mandatories, the bad briefs, the creative prisons, the Volkswagen ad, and the ancient sorrows of client feedback, let me give you the whole thing in its simplest form.

A great creative brief answers six questions:

1. Who are we trying to reach?
2. What tension, worry, desire, or pressure already lives in their world?
3. What one thing do we want them to believe?
4. Why should they believe it?
5. Who is the brand in this conversation?
6. What must be included, and what must not be changed?

That is it.

Everything else is elaboration, discipline, judgment, and pain.

The first question is the target customer. The second is the insight. The third is the benefit. The fourth is the reason why. The fifth is brand character. The sixth is mandatory executional elements and approval.

If you can answer those six questions clearly, you can write a useful creative brief. If you cannot answer them clearly, you are not ready to brief anyone, human or machine.

The rest of this essay explains why those questions matter, how to answer them, how they were taught to me, where they came from, how they fail, and how to turn them into an AI briefing system.

Because that is the new wrinkle. The creative brief is no longer just a way to align a client and an agency. It is becoming the strategic kernel of human-and-machine creative work.

A prompt asks the model to do something.

A brief tells the model what must remain true while it does it.

That distinction matters enormously.

---

The creative brief is an ad for the ad

Years ago Alison Tintle and I wrote the USCMO Green Book, which I have now posted here: Download the USCMO Green Book. The Green Book argued that the creative brief should be a single-page document whose clarity comes from simplification: one page, one business objective, one target customer, with signed approvals before agency work begins.

It also made the client’s obligation plain: if you do not do the thinking early, you will do it badly and expensively later, through rework, confusion, missed deadlines, and forgettable advertising.

Years ago, in that book and related materials, I kept returning to a simple construction: marketing is selling, advertising is the selling conversation, and advertising strategy is the selling proposition. What I would add now is this: the creative brief is the first disciplined, usable form of that proposition.

That is why the brief matters so much. It is not paperwork. It is not administration. It is not meeting minutes. It is the point at which the client stops admiring complexity and starts doing the harder and more valuable work of choosing.

Before the agency can make advertising, the brief has to sell the assignment to the people who will create it. It has to make the audience legible, the problem human, the promise sharp, and the reward of solving that problem feel real.

Now, increasingly, it has to do the same thing for a model.

That is why the brief is not becoming less important in the age of AI. It is becoming more important. AI does not replace the brief. AI wraps around the brief.

The brief remains the strategic center of gravity. The model can generate variations, routes, lines, drafts, concepts, counterarguments, syntheses, and provocations. But if the model is allowed to invent a new audience, a new benefit, a new reason to believe, or a new brand character every time it tries to be helpful, then you do not have creative exploration. You have strategic drift.

And strategic drift is not a new AI problem. It is an old client problem with a faster engine.

---

Table of Contents

• The short version: a great creative brief answers six questions
• The creative brief is an ad for the ad
• Another lineage that belongs here: Ken Roman, Jane Maas, and the Ogilvy school
• A word on P&G, Yves Ameline, and the copy workshop
• Many schools, same discipline
• What the one-page brief looks like
• How to complete each section
• A bad brief autopsy
• How to “improve” a Volkswagen ad until it dies
• Now turn the brief into a briefing system for AI
• A simple AI briefing stack
• Where briefs begin to die
• Do not put the creative behind bars
• The brief is only half the craft
• The Advertising Comment Organizer
• What happened to the training?
• Why this matters even more now
• Final thought

---

Another lineage that belongs here: Ken Roman, Jane Maas, and the Ogilvy school

There is another tradition behind my thinking on creative briefs that this essay would be remiss to ignore, and that is the Ogilvy school as carried forward by Ken Roman and Jane Maas.

When I started my first advertising agency, I made a very deliberate choice. There are always many plausible ways to do the work, many systems you can borrow, and many frameworks you can half-adopt and then mangle beyond recognition. I decided not to do that. I decided that How to Advertise would be our house Bible, and therefore How to Advertise was how we were going to advertise.

Everyone at the agency had a copy. Everyone read it. Everyone used it. I quoted from it constantly.

That choice mattered more than I realized at the time, because it taught me something I still believe very strongly: in advertising, as in many crafts, there is enormous value in choosing something coherent and standing for it. One of the great weaknesses of modern marketing culture is that people want the authority of a point of view without the discipline that point of view requires. They want the feeling of rigor without the inconvenience of commitment.

Good work usually comes from a more demanding place. It comes from saying, “This is the way we are going to think about brands, strategy, truth, tone, and persuasion,” and then holding yourself to that standard.

That is one reason Ken and Jane belong in this piece. Their work sits squarely inside the tradition this essay is trying to reclaim: brands, strategy, research, campaigns, target marketing, agency relationships, and, beneath all of that, a deeply practical respect for what advertising is meant to do.

Not self-expression.

Not internal theater.

Not the preservation of stakeholder feelings.

Persuasion.

And underneath that voice, of course, sits David Ogilvy. The Ogilvy view of the world has always appealed to me because it combines discipline with ambition. The work should be built on a real idea. Positioning matters enormously. The consumer deserves respect. Committees are dangerous. Advertising should sell.

Those are not antique opinions.

They are simply durable ones.

At the end of the day, people buy stories, people buy trust, and people buy into people. That is the part marketers forget when we lead with product. Product matters, of course, but product truth only becomes powerful when it enters a human story with enough force to change what someone believes, fears, hopes, or chooses.

That is not a soft statement. It is a hard one. If the product truth cannot enter a human story, the advertising will struggle. If the brief cannot describe that story, the creative team will be left to invent one. Sometimes they will. Often they will not. And in the AI age, the model will invent something too, but it may not be the thing you needed.

---

A word on P&G, Yves Ameline, and the copy workshop

Long before Microsoft, P&G had already taught me the deeper lesson anyway. When I think back to Yves Ameline and the copy workshop culture there, what I remember most is not mystique but standards: clarity, judgment, and simplicity.

You were expected to know what you meant.

You were expected to reduce.

You were expected to write in a way another human being could use.

I still remember my first copy workshop. I was a brand-new assistant brand manager, and it was a two-day off-site session at Ryerson. We would take turns being agency and client. If you were playing the agency role, your job was to present the storyboard persuasively against the brief. If you were playing the client role, your job was to critique the storyboard against the strategy.

Nothing about that sounds exotic, but in practice it was an extraordinary forcing mechanism for judgment.

You had to learn, quickly, how to distinguish between a strategic problem and an executional quibble. You had to learn what actually mattered in an ad, and what merely happened to be visible.

That was the gift Yves gave us. He gave us a language for advertising judgment.

Without that kind of training, people often look at advertising and, because they cannot really look at it strategically, they look at it executionally instead. They see a storyboard with six things happening and ask why the actor is wearing red, or why the kitchen looks modern, or whether the music feels slightly too jaunty.

They comment on the garnish because they do not know how to evaluate the meal.

Yves helped us move beyond that reflex. He gave us terms, categories, and confidence.

Irrelevant drama.

Video vampires.

Product truth.

Strategic fit.

He helped us understand that an ad can be lively without becoming distracting, and memorable without ceasing to sell.

That was not merely a workshop skill. At P&G it was a career skill. The company was ruthlessly serious about advertising judgment. People called it copy judgment, but what they really meant was the capacity to look at advertising and know what mattered.

Not everybody could do that.

Some people never learned.

And in the management culture of that time, a weakness of that kind could be fatal. It was a brutal system — a kind of corporate Hunger Games, really — and many good people disappeared quickly. You would make plans to have lunch with someone, go downstairs at noon, and find the desk already cleared out.

I learned a great deal there, and I do not romanticize the cruelty of it, but I do think the system understood something that many organizations no longer understand: if you are going to entrust marketers with brands, budgets, agencies, and mass persuasion, you had better teach them how to judge the work.

That training also taught a certain kind of compression. You can see it in those old P&G strategy sheets from the copy workshop. Their power lies in their refusal to wander. They reduce the whole question to promise, support, and character.

Charmin is squeezably soft. Charmin’s advertising should feel warm, friendly, and lighthearted. Butter Flavor Crisco offers superior baking performance versus margarine, with support and character stated in plain language.

No sludge.

No committee residue.

No attempt to preserve every stakeholder thought.

Just focus, clarity, and nerve.

The absence of that training is one reason so many briefs feel weak today. If you do not know how to judge advertising, you also tend not to know how to brief it. You do not know what to emphasize, what to omit, what kind of insight actually matters, or how to tell the difference between a real strategic proposition and an execution in search of a product.

I saw the opposite lesson very vividly later, during a month I spent as an intern at Leo Burnett under Martin Shewchuk, one of Canada’s great creative treasures. That month remains one of the happiest periods of my professional life.

One thing I saw there has stayed with me ever since. A creative had a storyboard with a gorilla in it — I am almost certain it was a gorilla — and he loved the idea so much that he pitched it to client after client after client, regardless of what they sold.

It was an execution in search of a product.

Maybe it was funny. Maybe it was even brilliant in isolation. But it was exactly backward. Good advertising does not begin with a favorite trick looking for somewhere to land. It begins with a product truth and a human truth, and then the execution grows from there.

Which is another way of saying that good creative briefs are not there to constrain the work. They are there to stop the work from floating away from the only truths that can make it powerful.

---

Many schools, same discipline

One reason I keep returning to these old documents is that they reveal something important. The discipline of the creative brief was never just one company’s paperwork fetish. Across the great agencies and great marketing companies, the forms differ, the language differs, and the culture differs, but the underlying grammar is remarkably consistent.

The Copy Workshop Strategy Seminar, based on work done for Apple Computer, collected strategy approaches from P&G, Y&R, GE, JWT, DDB/Needham, Leo Burnett, BBDO, Ogilvy & Mather, FCB, Tracy-Locke, Chiat/Day, and others. The striking thing is not how different these systems are. The striking thing is how often they circle the same truths.

School or model	What it teaches
P&G Copy Strategy Memo	State the basic benefit on which consumers are expected to choose the brand. Keep the strategy clear, simple, competitive, and free of executional considerations.
Y&R Creative Work Plan	Identify the key fact, the consumer problem advertising must solve, the advertising objective, the prospect, the promise, the reason why, and the mandatories.
GE Focus System	Focus first on the receiver, then on the proposition, and only then dramatize the proposition. Dramatization before focus is waste.
JWT Creative Brief	Define the opportunity or problem, the desired response, the audience, the key response, supporting information, brand personality, and practical constraints.
DDB/Needham R.O.I.	Great advertising needs relevance, originality, and impact. Promise and support are central decisions, and personality gives the brand distinction.
Leo Burnett Strategy Worksheet	Move from target audience and current belief to support, focus of sale, proposition, and desired belief.
BBDO	Advertising must marry the product image and the user image. Product truths become powerful when joined to the person who experiences them.
Tracy-Locke Six Elements	Target audience, user benefits, reason why, brand character, focus of sale, and tone. The focus of sale is the principal basis on which the sale stands or falls.

That is a lot of vocabulary to arrive, more or less, at the same place.

Who are we talking to?

What problem or opportunity matters?

What promise are we making?

Why should anyone believe it?

What kind of brand is speaking?

What must the creative team respect?

Different schools. Different dialects. Same grammar.

The forms matter because they create procedure. Jobs are initiated. Work plans are written. Strategies are approved. Creative is evaluated against the strategy rather than against personal taste. When things go well, the documents are studied. When things go badly, the documents help reveal where the thinking broke.

But the forms are still only pieces of paper.

Smart people have to use them smartly.

---

What the one-page brief looks like

Over time I have come to prefer a form that is even simpler than the formal Green Book template. The Green Book had the right instincts: one page, a clear objective, one target, clear strategy, project budget, mandatory elements, and signed approvals. My own current version is really just a further compression of that logic.

It assumes the business objective has already been settled before this page is written, and it forces the brief itself to become brutally usable.

And again, the old P&G sheets are a useful reminder of what brutal simplicity looks like. They do not try to sound sophisticated. They try to be clear.

That is the standard.

CREATIVE BRIEF

Project Name: _________________________________________________

Project Due Date: _____________________________________________

Project Budget: _______________________________________________

Type of Project: ______________________________________________

Target Customer:
_______________________________________________________________
_______________________________________________________________

Target Customer Insight:
_______________________________________________________________
_______________________________________________________________

Benefit:
_______________________________________________________________
_______________________________________________________________

Reason Why:
_______________________________________________________________
_______________________________________________________________

Brand Character:
_______________________________________________________________
_______________________________________________________________

Mandatory Executional Elements:
_______________________________________________________________
_______________________________________________________________

Approved by: _________________________________________________

I like this form because it leaves nowhere to hide. It clears the obvious operational questions immediately and then forces the strategic core into a shape small enough that evasion becomes difficult.

There is one important caveat. If the business objective is still fuzzy, stop. Do not proceed. A brief needs one business objective, and if there is more than one business objective there should be more than one brief.

Awareness is not a business goal.

Vagueness is not strategy.

“We need to tell everyone everything” is not an objective. It is a confession.

---

How to complete each section

Project Name

The Project Name should be descriptive rather than cute, because this is not the place for internal whimsy. The project name exists so people can understand what this is when they see it in a folder, calendar invite, approval chain, or AI workspace three months from now.

Do not call it “Project Phoenix” unless a phoenix is actually involved.

Project Due Date

The Project Due Date should be real rather than aspirational, because timing shapes ambition, production, and the amount of exploration an agency can responsibly undertake. A campaign that needs to be in market in six weeks is a different assignment from one that has six months to breathe.

Time is not a footnote. Time is a creative constraint.

Project Budget

The Project Budget belongs up front because budget is not an embarrassing accounting detail to be revealed later. It tells the agency what game is actually being played.

Do not brief an epic and then pay for a pamphlet.

Type of Project

The Type of Project should be equally explicit. A launch film is not a landing page, a landing page is not a nurture stream, and a nurture stream is not a field-event kit. Creative people think differently when they know the form of the ask.

AI needs this too. One of the most common mistakes people make with models is to ask for “ideas” without saying what kind of artifact those ideas are supposed to become. A keynote opening, a product demo script, a paid social campaign, and a customer email all require different thinking.

“Give me ideas” is not an assignment. It is a cry for help.

Target Customer

The first genuinely difficult line is Target Customer, and this is where many briefs quietly begin to die.

If you are trying to reach a buyer, a user, an influencer, a procurement person, and an internal stakeholder all in the same document, you are not being nuanced. You are avoiding choice.

One brief should have one target customer. If there are genuinely different customers with genuinely different jobs to be done, there should be separate briefs.

The art here is to be concise without becoming vague. Give the creative team enough to understand the person, the pressure under which that person is operating, the role that person inhabits, and the relevant business or life situation, but do not turn the brief into a novella.

A useful target customer line might say:

Mid-market CIOs who are being asked to enable AI experimentation across the company faster than their security, compliance, and governance models can comfortably support.

That is much better than:

IT decision-makers.

The first one contains a person, a situation, and a pressure. The second one contains a database field.

Target Customer Insight

The next line, Target Customer Insight, is where marketers often commit a different kind of sin. They either write something so generic that it could apply to anyone, or they quietly bury the benefit there because they have not had the nerve to state it openly in the proper place.

An insight is not trivia, and it is not the surface description of the problem. It lives one level deeper, at the level of implication.

What, exactly, worries the target?

What is the target trying to protect?

What social, parental, professional, financial, operational, or identity consequence sits beneath the visible problem?

The difference matters enormously. To say, for example, that my children go to school in stained clothes because I cannot get them clean is useful context, but it is not yet an insight. It is still sitting on the surface. The insight is that other parents will think I am a bad parent, and my children will think I do not love them.

That is the level at which the brief begins to acquire voltage.

The insight should make the benefit necessary.

It should not duplicate the benefit.

Benefit

Benefit is the line around which the entire brief turns. This is the promise. This is the proposition. This is the one thought the work is supposed to leave in the customer’s mind.

Be disciplined here.

In B2B, most benefits eventually resolve into three very old outcomes:

• Help me increase revenue.
• Help me decrease cost.
• Help me manage risk.

You can dress those outcomes in more artful language, and often you should, but you should not kid yourself about what lies beneath. The point of a benefit is not that it sounds strategic. The point of a benefit is that it names something someone worries about whether they know about your product or not.

It is something already present in life.

It is already part of the person’s day, part of the person’s sleep, part of the person’s anxiety, part of the person’s ambition.

That is why these supposedly “business” benefits are so emotional. Increased revenue is emotional because it means growth, praise, bonus, status, and career momentum. Decreased cost is emotional because it means control, credibility, relief, and the ability to walk into the operating review without dread. Managed risk is emotional because failure is emotional.

If the risk lands badly enough, you are halfway into a country song.

So do not make the mistake of thinking that business benefits lack feeling.

They are packed with feeling.

The real discipline is to express the benefit in terms of something the customer worries about whether they know about your product or not.

Usually not.

One of the classic examples is the old Listerine campaign built around “Always a bridesmaid, never a bride.” By contemporary standards the execution would be completely unacceptable, but strategically it was brilliant. The woman in the ad was not sitting around wishing for antiseptic mouthwash. She was worrying about why she could not get married. Listerine entered the story as the hidden solution to a problem that already carried emotional force.

Insurance works the same way. What happens to my family when I am gone? Will I have enough money when I retire? Those are the worries. The product arrives second.

That is why marketers need to be careful not to sneer at simple benefits, domestic benefits, or benefits that seem too ordinary to deserve respect. The ordinary often carries the deepest emotional charge.

When I was a boy, my mother was running a household at full speed with three children and a husband who was hardly ever home. Some mornings the laundry was still in motion as we were getting ready for school. I remember shirts coming straight out of the dryer at the last possible minute.

Only years later did my mother tell me that those mornings felt to her like panic, like proof that the day had already escaped her control before it had truly begun. But that was not what I felt as a child.

What I felt was a warm shirt fresh from the dryer, and that warm shirt felt like a hug from my Mom. The warmth of that shirt carried me out of the house on my way to school. For her, the moment may have felt like failure. For me, it felt like love.

The same lesson showed up in our kitchen. My mother used to wash the top of a can before opening it. One day I asked why. Her first answer was not the practical one. Her first answer was, “Because I love you.”

That was the benefit.

Only then did she explain the reason why: can lids are dirty, factories are dirty, and once the opener punctures the lid whatever sits on top can wind up inside.

That was the proof.

Years later I saw her open a can without washing it first, and it landed with surprising force. In my child’s logic she had not merely skipped a hygiene step. She had withdrawn the promise. She had killed the benefit.

For the record, I always wash a can before opening it, even if it’s just for me.

Or my dog.

That is how emotion actually works in life. It hides inside ordinary acts. It hides inside shirts, soup cans, budgets, margins, payroll, security, school clothes, and mortgages.

Which is why a good benefit is never merely functional. If it matters at all, it already has emotional weight. People buy stories. People buy trust. People buy into people. Even the most prosaic benefit must eventually connect to some human truth, or else it will never travel very far.

I still like to force this line to begin with the words To convince, because that phrasing has a way of smuggling discipline into the room.

To convince whom of what?

That is almost always the real question anyway, so you might as well face it directly. If you cannot answer it cleanly, you are not ready to brief.

Reason Why

Reason Why is the functional support. It answers the question the customer will ask immediately after the promise is made, namely: why should I believe you?

This is where proof belongs, but only the proof that actually supports the chosen promise. Not every feature, not every claim sales has ever collected, and not every talking point to which somebody in the building feels emotionally attached.

The reason why exists to support the benefit in plain, simple language. It is not there to restart the brief with three new strategies.

That little word supports does a great deal of work.

Brand Character

Brand Character is where you tell the agency who the brand is when it walks into the room.

This should not become adjective salad. It should be a judgment about personality. Bounty is practical and indispensable. Charmin is warm, friendly, and lighthearted. Cheer is engaging, full-of-life, and up-to-date. Butter Flavor Crisco inspires confidence and pride.

That is what brand character is supposed to do. It tells the creative team how the brand sounds, behaves, and carries itself.

AI needs this even more than humans do. A good creative team can often infer tone from the brand, the category, the history, the audience, and the client’s body language in the room. A model cannot read the room unless you write the room into the system.

If you do not tell it who the brand is, it will default to the median voice of the internet, which is usually cheerful, competent, bland, and faintly dead inside.

Mandatory Executional Elements

The final two lines matter more than people think. Mandatory Executional Elements should be brutally short and genuinely mandatory: legal language, a required claim, a logo, a URL, a product shot, an offer, and similar items.

The key word is not elements.

The key word is mandatory.

The moment this section becomes a parking lot for stakeholder anxiety, the creative is behind bars.

Approved by

And Approved by is not ceremonial. Signed approval is simply the adult act of owning the strategy before the work begins. It protects against the extremely common organizational habit of pretending, three weeks later, that nobody ever agreed to the very thing the agency was briefed to make.

In an AI system, approval matters in a slightly different way. It tells you when to freeze the strategic core. The model can help explore, sharpen, challenge, and improve a draft brief before approval. But once the brief is approved, the approved strategy should become the ground truth.

Do not let the model casually keep rewriting the premise unless you have specifically asked it to help rewrite the premise.

---

A bad brief autopsy

It may help to see the difference between a weak brief and a strong one.

Let us imagine a fictional product: an AI governance platform for mid-market companies. It helps teams approve, monitor, and control AI usage across the organization without slowing everyone to a crawl.

Here is the kind of bad brief that gets written every day:

Brief Section	Weak Version
Target Customer	IT leaders, security teams, compliance teams, business users, executives, and employees using AI.
Target Customer Insight	Companies want to use AI safely and productively.
Benefit	Our platform improves AI governance, productivity, compliance, security, collaboration, and innovation.
Reason Why	We have dashboards, workflows, integrations, policies, analytics, alerts, reports, and role-based access.
Brand Character	Innovative, trusted, human, bold, simple, enterprise-grade, friendly, secure, modern, and approachable.
Mandatories	Logo, URL, AI message, security message, productivity message, compliance message, innovation message, customer quote, Gartner mention, product screenshot, partner logo, and “responsible AI” language.

That brief is not useless because it lacks information. It is useless because it lacks judgment.

It contains many plausible inputs, but it has made almost no strategic choices. The audience is plural. The benefit is plural. The support is a product inventory. The brand character is a salad. The mandatories are a hostage situation.

If you handed that brief to a creative team, they would have to do the actual strategy before they could do the creative. If you handed that brief to a model, the model would politely absorb the indecision and produce a campaign that sounded like every AI governance campaign you have ever seen.

Now here is a better version:

Brief Section	Stronger Version
Target Customer	Mid-market CIOs who are being asked to enable AI experimentation across the company faster than their security and governance models can comfortably support.
Target Customer Insight	They do not want to be the person who says no to AI, but they are terrified of becoming the person blamed for the AI breach everyone else pressured them to allow.
Benefit	To convince CIOs that they can say yes to AI adoption without losing control of risk.
Reason Why	The platform gives IT one place to approve AI tools, enforce usage policies, monitor adoption, and document compliance across teams.
Brand Character	Calm, practical, and authoritative — the grown-up in the AI room.
Mandatories	Include approved product name, URL, “AI governance without gridlock” claim, and required security disclaimer.

Notice what changed.

The stronger brief did not get longer. It got sharper.

It chose a customer. It named a pressure. It turned that pressure into a promise. It selected the proof that supports that promise. It gave the brand a usable character. It cut the mandatories down to the things that are actually mandatory.

The weaker brief preserved inputs.

The stronger brief made choices.

That is the whole game.

---

How to “improve” a Volkswagen ad until it dies

There is a wonderful and painful piece called Nine Ways to Improve an Ad, written by Fred Manley, then V.P., Creative Director at BBDO San Francisco, with illustrations by Hal Riney. It first appeared in CA in the July/August 1963 issue, and it may be one of the best demonstrations ever made of how advertising gets destroyed by apparently sensible improvements.

The piece begins with the classic Volkswagen “Think small.” ad. Then Manley reconstructs what the ad might have become had it followed the sensible rules that guide so much advertising.

First, do not use a negative headline. “Think small.” becomes “Think BIG!”

Then make the thought more positive and more direct.

Then show people enjoying the product.

Then give prominent display to the product logo.

Then make sure the reader knows where to buy the product.

Then add dealer information.

Then add excitement.

Then add convenience.

Then add credit.

Then add coffee.

Then add the dealer names.

And by the time the process is complete, the great ad has been fixed into a shrieking carnival of helpfulness.

That is the perfect verb, by the way. Not ruined. Fixed.

The original Volkswagen work worked because it was strategically brave and executionally restrained. “Think small.” was not cleverness for its own sake. It was the dramatization of a proposition. Volkswagen could not win by pretending to be a big American car. It won by making smallness a virtue.

The discipline of the ad is what makes it beautiful. It does not shout. It does not over-explain. It does not give every stakeholder a little patch of real estate. It does not put the product under a disco ball and beg for attention.

It trusts the idea.

What the fake “improvements” reveal is how easily people destroy that trust. They do not destroy it because they hate creativity. They destroy it because they are uncomfortable with the discipline that creativity requires.

They want to make sure the ad says everything.

They want to make sure the customer cannot miss the point.

They want to make sure the legal team is happy.

They want to make sure the product team sees its favorite feature.

They want to make sure the sales team sees its favorite proof point.

They want to make sure the executive sees the logo.

They want to make sure nobody can object.

And when nobody can object, nobody can remember it either.

This is why the creative brief matters. A great brief gives the team permission to protect the idea from well-intentioned erosion. It says: this is the customer, this is the tension, this is the promise, this is the proof, this is the character, and these are the real constraints.

Everything else is not a mandatory.

It is anxiety.

---

Now turn the brief into a briefing system for AI

Once you understand the periodic table of a great creative brief, the next question becomes obvious: where does AI fit?

The answer is that AI does not replace the brief.

AI wraps around the brief.

That distinction matters. A creative brief is still the strategic core. It still defines the audience, the insight, the benefit, the reason why, the brand character, and the mandatory elements. But in an AI environment the brief becomes one layer inside a larger system.

The old brief does not disappear.

It becomes the strategic center of gravity.

A useful way to think about this system is as a stack.

1. The Brand Layer

The first layer is the Brand Layer. This is the durable material that should remain true across campaigns: voice, values, vocabulary, approved claims, forbidden claims, legal guardrails, visual codes, tone boundaries, and category landmines.

This layer is not campaign-specific. It is the enduring operating context of the brand.

It tells the model what kind of company it is working for before it starts generating anything.

2. The Creative Brief Layer

The second layer is the Creative Brief Layer. This is the one-page brief you have just written: target customer, target customer insight, benefit, reason why, brand character, and mandatory elements.

This layer is the strategic center. Freeze it. Do not let the model casually invent a new audience, a new benefit, or a more “interesting” strategy.

That is not creativity.

That is drift.

3. The Assignment Layer

The third layer is the Assignment Layer. This is where you tell the model what job it is solving right now.

Is the deliverable a landing page, a launch film, a manifesto, a keynote opening, a paid social campaign, a set of email subject lines, a customer story, a trade-show concept, or a sales deck?

What channel?

What format?

What length?

How many routes?

What specific response should the work create?

One of the most common errors in AI prompting is to leave this layer fuzzy and hope the model will infer the task. Do not make it infer.

Tell it.

4. The Quality Layer

The fourth layer is the Quality Layer. This is where you define what good looks like.

Should the work feel bold or restrained? Premium or plainspoken? Familiar or surprising? Should it avoid category clichés? Should it produce three distinct routes or one route explored in depth? Should it optimize for memorability, persuasion, elegance, speed, clarity, emotional force, or provocation?

If you do not define the standard, the model will invent one.

And the model’s invented standard will usually be “reasonable.”

Reasonable is where ideas go to nap.

5. The Evaluation Layer

The fifth layer is the Evaluation Layer. This is the layer many people forget, and it may be the most important.

Before the model generates anything, make it restate the strategy in one sentence.

Who is the audience?

What does that audience worry about?

What one thing are we trying to convince that audience to believe?

Why should that audience believe it?

If the model cannot answer those questions clearly, then it should not yet be generating. It should be asking for clarification.

That is the key shift.

You do not want the model inventing new audiences, new benefits, or new strategy in the name of helpfulness. You want disciplined variation in execution while the strategic center remains fixed.

Let the human do the choosing.

Let the machine do the exploring.

Put a little more bluntly: never let the model rewrite the brief unless you have explicitly asked it to do so.

And one more point matters enormously: the brief belongs inside the AI system, not off to the side. The flow should be simple.

1. Write and approve the brief.
2. Freeze the strategic core.
3. Wrap the brand layer, assignment layer, quality layer, and evaluation layer around it.
4. Ask the model to restate the strategy before it generates.
5. Generate multiple routes.
6. Evaluate the work against the brief rather than against the mood in the room.

That is where the stack earns its keep. It gives you a place for each kind of truth, and it stops the machine from wandering.

This is where a great creative brief stops being a historical artifact and becomes a modern operating system.

---

A simple AI briefing stack

Here is a practical starting point. This is not the only way to do it, but it is a good discipline.

You are a senior creative partner working from an approved creative brief.

Do not invent a new audience, a new benefit, or a new strategy.

If the brief is ambiguous, ask questions before generating.

A prompt asks you to do something.
A brief tells you what must remain true while you do it.

BRAND LAYER

Brand voice:
[Insert durable brand voice guidance.]

Brand values:
[Insert durable brand values.]

Approved vocabulary:
[Insert words, phrases, naming conventions, and claims the brand uses.]

Forbidden vocabulary:
[Insert words, phrases, claims, or category clichés the brand avoids.]

Claims allowed:
[Insert claims approved for use.]

Claims forbidden:
[Insert claims that must not be made.]

Legal, regulatory, or category guardrails:
[Insert legal requirements, disclaimers, market constraints, and category landmines.]

CREATIVE BRIEF LAYER

Target customer:
[Insert one target customer.]

Target customer insight:
[Insert the pressure, worry, desire, or consequence beneath the visible problem.]

Benefit:
[Begin with “To convince...” and state the one thing we want the audience to believe.]

Reason why:
[Insert the proof that supports the benefit.]

Brand character:
[Insert who the brand is when it walks into the room.]

Mandatory executional elements:
[Insert only the genuinely mandatory elements.]

ASSIGNMENT LAYER

Deliverable:
[Insert the artifact: landing page, launch film, email, keynote opening, paid social, sales deck, etc.]

Channel:
[Insert channel or environment.]

Audience sophistication:
[Insert what the audience already knows or believes.]

Desired response:
[Insert what the audience should think, feel, do, question, or remember.]

Length and format:
[Insert length, format, structural requirements, or production constraints.]

Number of routes requested:
[Insert number of distinct creative directions.]

QUALITY LAYER

What good looks like:
[Define the standard.]

What to avoid:
[Define category clichés, tonal mistakes, strategic traps, and executional traps.]

Reference tones or models:
[Optional: insert references, but do not ask the model to imitate copyrighted text or living authors directly.]

How bold or conservative to be:
[Set the ambition level.]

EVALUATION LAYER

Before generating, restate the strategy in one sentence.

Then answer:
1. Who is the audience?
2. What does that audience worry about?
3. What do we want the audience to believe?
4. Why should the audience believe it?
5. What must remain true across every route?

If the strategy is unclear, ask clarifying questions before generating.

Then produce:
- 3 distinct creative routes
- A short rationale for each route
- A self-check against the approved brief
- A note on where each route is strongest and weakest

The important thing is the principle: the brief is the strategic core, and the AI system is the operating wrapper around it.

This has another useful effect. It makes AI output easier to evaluate. Instead of asking, “Do we like this?” you can ask better questions:

• Did it stay with the approved target customer?
• Did it dramatize the insight?
• Did it make the benefit easier to believe?
• Did the proof support the promise?
• Did the brand character come through?
• Did it obey the real mandatories without becoming imprisoned by them?

Those are better questions because they return the conversation to strategy. The enemy of good creative evaluation is not disagreement. Disagreement can be useful. The enemy is preference masquerading as judgment.

AI will make that problem worse unless we get more disciplined. It will be very easy to generate twenty versions of something and then let everyone pick the one that best reflects their private taste. That is not creative development. That is a buffet.

The brief is what keeps the buffet from becoming the strategy.

---

Where briefs begin to die

Once you see the brief as both a human document and a machine-readable strategy, its failures become even easier to recognize.

A weak marketer can turn even a strong template into mush.

A strong marketer can use a plain template to create extraordinary clarity.

The difference is judgment.

That is why bad briefs become depressingly easy to recognize after you have read enough of them. They usually die in one of three places:

1. The audience is plural.
2. The promise is plural.
3. The courage required to state the real thing plainly has failed somewhere along the line.

The multiple-audience brief is almost always a fear document. Somebody is afraid that narrowing the audience will insult another stakeholder, leave money on the table, or make the campaign seem too small for the business. So the brief starts trying to speak to everybody at once.

The result is that it stops speaking clearly to anybody in particular.

Something similar happens with benefits. Instead of choosing the one thing the work most needs to do, the author begins collecting good things the product can do and stacking them together in the hope that abundance will be mistaken for strategic richness.

It rarely is.

What it usually produces is a pile of half-articulated promises that no creative team could dramatize with any real force.

The third failure is subtler, and it appears when the author is nervous about saying the hard thing directly. That is when the real benefit gets buried in the insight, hinted at in the mandatories, implied somewhere in the tone language, or softened into a paragraph of corporate vapor.

This is one reason I keep coming back to muddy writing. Muddy writing is usually just muddy thinking made visible. When the choices have actually been made, the prose has an uncanny tendency to become clear.

That image remains one of the truest things ever drawn about advertising. A little contribution here, a little contribution there, and before long the idea is still technically alive but no longer recognizably itself.

AI adds a new version of this problem. The model will often try to be helpful by adding what you forgot. That can be useful during exploration. It can be dangerous during execution.

If the model adds a new audience, it has changed the brief.

If it adds a new promise, it has changed the brief.

If it turns the proof into a second strategy, it has changed the brief.

If it takes a clear brand character and turns it into generic professional warmth, it has changed the brief.

Do not blame the model for that. Blame the system that failed to tell the model what must remain true.

---

Do not put the creative behind bars

There is another way briefs go bad, and it is slightly different from committee thumbprints. In this version, the brief becomes a jail cell built from rules: brand rules, tone rules, legal rules, design rules, format rules, and inherited rules from campaigns nobody even remembers.

Soon enough the brief reads less like a provocation and more like parole conditions.

To be fair, rules matter. Legal matters. Product truth matters. Real constraints matter.

But there is a vast difference between helpful constraint and creative incarceration.

Rules are supposed to protect the truth, not replace thought. The job of the brief is to tell the creative team where the walls actually are, not to build a maze inside the cell.

This is especially important with AI, because models are literal-minded in ways humans are not. If you give the model a list of twenty constraints, it may treat all twenty as equally important. It does not know which are real legal requirements, which are brand preferences, which are inherited superstitions, and which are simply someone’s old pet peeve that has somehow become canon.

That is why the brief should distinguish between:

• True mandatories: things that must be included or avoided.
• Strategic guidance: things that help shape the work.
• Preferences: things someone likes but which are not binding.
• Superstitions: things the organization keeps repeating because nobody remembers why they exist.

A good creative brief is not a rule dump. It is a judgment document.

---

The brief is only half the craft

Even a very good brief can still be ruined by weak client behavior, which is one reason I have never thought of briefing as a purely clerical exercise. The brief is the beginning of the work, not the end of the client’s responsibility.

A creative brief is where the client chooses the strategy. A creative presentation is where the client proves whether it meant what it chose.

That distinction matters. Many clients can approve a brief in principle and then abandon it the first time an idea makes them nervous. They nod at focus in the abstract, then ask the agency to add three more messages, make the logo bigger, soften the tension, clarify the joke, satisfy the sales team, reassure legal, flatter the product team, and make sure no stakeholder feels unseen.

That is how advertising gets fixed to death.

It is also how agentic advertising will get fixed to death. The only difference is speed. A human agency may take a week to come back with a compromised revision. An AI agent can produce seventeen compromised revisions before lunch.

This is where the old Advertising Comment Organizer still feels wonderfully modern. It gives clients a way to respond to creative work without turning every presentation into a free-association exercise. Better still, it adapts beautifully to the AI age. The same structure that helps you comment on an agency’s storyboard can help you comment on an agent’s campaign route, landing page, email sequence, film script, or social concept.

The point is not to make clients timid. Quite the opposite. The point is to make clients useful.

---

The Advertising Comment Organizer

This is a one-page artifact you can use before giving feedback to an agency, an internal team, a freelancer, or an AI agent. Print it, paste it into a shared document, or use it as the evaluation prompt inside an agentic workflow.

The power of this organizer is that it slows the client down just enough to become precise. It asks the client to understand before reacting, evaluate before commenting, and communicate before tinkering.

That sequence matters. Bad comments usually happen when people skip straight from discomfort to prescription.

“Make it warmer.”

“Can the logo come in earlier?”

“What if the headline were more direct?”

“Can we make it feel more premium?”

“Can the agent generate twenty more options?”

Maybe those are useful comments. Often they are not. The organizer forces a better question first: what problem am I trying to solve with this comment?

If the work is off strategy, say so. If the selling line is weak, say so. If the end-result benefit is not visible, say so. If the drama is distracting from the product truth, say so. If the comment is merely a preference, have the discipline to know that too.

This distinction becomes even more important with agentic advertising, because AI makes reaction feel productive. You can keep asking for more versions forever. You can generate endlessly. You can tweak tone, length, audience, format, structure, metaphor, headline style, proof points, calls to action, and mandatories until the original strategy is no longer visible.

This is why the client’s craft still matters. Someone still has to protect the strategy. Someone still has to judge the work against the approved brief. Someone still has to say: this is the route, this is why it works, this is what must be fixed, and this is what must not be touched.

The same principle sits behind the old Ten Steps to Becoming a Great Client. They are not a side topic. They are the behavioral counterpart to the brief itself.

1. Focus on the unique advantage

Find the one thing that is truly different and truly special. If you had three seconds to make a pitch and your dinner depended on the sale, what would you say?

That forcing mechanism is healthy because it strips away corporate noise and gets you back to the one thing that matters most.

2. Stay big picture in creative presentations

Most clients destroy work by commenting on the tenth-most-important thing first. The right questions are bigger: what is the selling line, what is the key visual, what is the dramatic flow, and what will the customer remember after the ad has ended?

If you start with the garnish, do not be surprised when the meal suffers.

3. Be honest with the agency

Act like a real person with real reactions. If you love something, say so. If you hate something, say so. If you are unsure, say that.

Agencies cannot work with strategic theater or manipulated ambiguity. They can work with clear, adult honesty.

4. Approve or kill in the third meeting

Overmeeting is one of the great silent killers of advertising. Endless rounds of tiny revisions demotivate the agency, flatten the work, and create the illusion of progress without any real advance.

Decide.

Move.

5. Take risks

Nothing motivates like responsibility. Taking an agency recommendation unchanged is an enormous vote of confidence, and great agencies respond to that trust. They will keep refining the work without needing you to rebuild it line by line.

6. Demand excellence

Benchmark against the best, not merely against what you produced last quarter. High standards make better work possible.

What crushes people is not excellence.

What crushes people is nitpicking masquerading as excellence.

7. Set clear objectives

Your agency cannot read your mind. Your agent cannot read your mind either. If you have a strategic conviction, say it plainly. Hidden agendas and telepathic expectations are a miserable way to work, and they lead directly to the sentence every client eventually regrets:

8. Let the agency make money

This point is more important than many marketers like to admit. An underpaid agency is a distracted agency. If you keep an axe over its head, it will spend time protecting its own business instead of growing yours.

The AI version of this is different but related: do not mistake cheap iteration for good process. Just because the machine can generate more does not mean the team should evaluate more. Abundance without judgment is not leverage. It is clutter.

9. Stay current

Do not drift away from the world your customers actually inhabit. Read strange things. Go places. Stay interested. Stay porous.

Creative work dies quickly in the hands of marketers who have become too insulated to recognize living culture when they see it.

10. Have fun

Advertising is serious work, but it is also joyful work when done well. Enthusiasm matters. Generosity matters. Clients who love the game tend to get better work than clients who treat every meeting as a compliance hearing.

There is a special kind of tragedy in organizations that put great effort into sharpening a brief and then immediately fall apart in the response phase. Suddenly the conversation is no longer about the selling line, the key visual, or the dramatic effect a customer will actually remember.

The conversation is about fonts, pet peeves, inherited preferences, hidden agendas, or the peculiar wish to revisit a strategy that was supposedly agreed three meetings earlier.

The longer I do this, the more I believe that a creative brief is not just a form. It is a behavioral contract. It says that these are the adults, this is the strategy, this is the proposition, and this is the terrain on which the work will now be judged.

The same will be true when the work is made with AI. Someone still has to behave like the client. Someone still has to protect the strategy, evaluate the routes, make the choices, and resist the temptation to keep asking the model for “just one more version” because choosing is uncomfortable.

---

What happened to the training?

This, more than anything else, is the question that keeps nagging at me.

What happened to the training?

I do not mean some abstract training budget, a library of templates in a shared drive, or a quarterly lunch-and-learn during which everyone nods politely and then goes back to producing the same documents that existed beforehand.

I mean actual training: editing, distillation, rewriting, and somebody telling you that your second audience is a dodge, your insight is just an observation, or your reason why is trying to become a second strategy.

I mean somebody telling you that the problem is not the words on the page but the fact that you have not yet chosen.

Most younger marketers I work with today would not last a week in the old P&G copy-workshop culture, not because they are lazy, but because they have not been formed by that discipline.

They have been taught how to gather inputs, align stakeholders, populate templates, and socialize documents. They have not always been taught how to distill. They have not been forced to learn that a page can get better by getting shorter, or that a strategy can get stronger by becoming more severe.

The result is that many briefs now read like unresolved meetings: overfull, politically careful, strategically hesitant, and oddly frightened of focus.

That is a shame, because the whole point of a creative brief is to perform that reduction.

It is not there to preserve complexity.

It is there to turn complexity into direction.

This is also why AI training cannot simply be “prompt training.” Prompt training is useful, but if it is not built on marketing judgment, it becomes another surface skill. People learn how to ask for outputs without learning how to recognize whether those outputs are strategically sound.

That is not enough.

The next generation of marketers does not only need to learn how to prompt. They need to learn how to brief, how to judge, how to choose, and how to protect a proposition from drift.

Otherwise we will create a generation of very fast marketers who can produce limitless amounts of work with very little idea whether any of it should exist.

---

Why this matters even more now

All of this would matter even if AI had never shown up, because advertising has always depended on clear propositions, clear audiences, and clear reasons to believe.

AI, however, has made the whole issue much less forgiving.

A model will do exactly what a half-decent agency does when the strategy is muddy: it will average the confusion beautifully.

That is why I keep returning to the idea that a good creative brief is becoming, in practice, a good prompt. If you give an LLM a sharp audience, a real tension, a clear benefit, believable support, and a usable brand character, it can be extraordinarily productive.

If you give it multiple audiences, multiple benefits, a hidden objective, contradictory instructions, and a fog bank of stakeholder language, it will faithfully synthesize the mess and hand it back in polished form.

There is a temptation, when people talk about AI, to imagine that the machine will somehow rescue the weakness of the humans operating it. I do not think that is how this will work.

The machine will reward clarity where clarity exists and amplify vagueness where vagueness exists.

That is why I think this old topic has become newly urgent. The creative brief has turned out not to be a relic of a slower era, but one of the most important leverage points in a faster one.

This is not a museum piece.

It is not a sentimental look backward.

It is not an argument for going back to some imagined golden age.

It is an argument for taking one of the oldest disciplines in marketing and using it to build something new.

Learn the periodic table. Learn the elements. Learn how the great briefs of the past were built. Then take that knowledge and use it to create the next generation of prompts, systems, campaigns, and creative work.

Maybe, just maybe, the next campaign that feels as startling and inevitable as the great Volkswagen work will not come from treating AI as magic. Maybe it will come from briefing it so brilliantly, so simply, and so truthfully that something brilliant, simple, and true finally comes back.

And maybe the future of creative work will belong not to the people who generate the most options, but to the people who know what must remain true across all of them.

---

Final thought

The creative brief is not where we avoid hard choices.

It is where we make them.

It is the client’s contribution, the first disciplined form of the selling proposition, and the point at which strategy either becomes usable or dissolves into mush.

For years that mattered because agencies could only make great work from clear inputs. Now it matters even more, because machines cannot save us from vagueness. They can only scale it.

So when you find yourself staring at a brief with multiple audiences, multiple benefits, a buried strategy, and a list of mandatory elements long enough to frighten a prisoner, resist the modern temptation to add another paragraph and call the result nuance.

Go back and do the thing the brief was supposed to do in the first place.

Choose the customer.

Choose the tension.

Choose the promise.

Choose the proof.

Choose the character.

Then sign it as though you mean it.

Because whether you are briefing an agency, an internal team, a freelancer, or an LLM, the principle is still exactly the same: