Use AI for the research behind a campaign brief, and make the decisions in it yourself. A campaign brief records decisions the business has already made and passes them to the people doing the work: the agency, the creative team or the media team. AI is good at gathering and sorting the evidence those decisions rest on. It's poor at making the decisions, because a model leans towards what the category already does and accepts the first answer that sounds reasonable.

The temptation is real. Writing a campaign brief takes time, and a model will produce a complete-looking one in under a minute.

What AI Does Well on a Campaign Brief

AI is useful for the research that sits underneath a campaign brief. It can analyse data from several sources at once, build a picture of the audience from interviews and reviews, and list the claims competitors in your category make.

In each case the model gathers and sorts the material faster than you can, and you decide what it means. For example, a model can turn three years of win and loss notes into a list of the reasons deals were lost in a few minutes. A person might take a day. Choosing which of those reasons the campaign should address is a decision about where the business spends its money, and that decision stays with you.

The list of competitor claims is the most overlooked of these uses. Ask a model what every competitor in your category says about itself, and you get a summary of the messages buyers already hear. Read it before you write the proposition, so you know which claims are already taken.

Why AI Writes an Average Campaign Brief

A model trained on what has already been written produces more of the same. It leans towards the conventions of the category, and you write a campaign brief to move away from them. If your brief repeats what every competitor says, the agency or creative team has no distinct idea to work from, and the work that comes back will be just as ordinary.

The bigger problem is that a model stops at the first answer that makes sense. Good briefing works the other way. You ask why the business wants this campaign, then why that matters, then why now. The useful answer usually comes three or four questions in, often after someone has had to admit something uncomfortable. A model has no reason to keep asking, because it has nothing riding on the answer.

A polished draft is also harder to challenge than a rough one. When a brief arrives formatted, fluent and complete, readers assume every section reflects a decision someone made. Nobody asks the question they would have asked if the section had been left blank.

The Decisions AI Can't Make for You

Some parts of a campaign brief are commercial decisions, and someone in the business has to carry the consequences of each one:

• what the work has to return, and by when
• which markets and customers are in scope, and which are out
• why a customer would choose you over the alternatives, including doing nothing
• what the business will spend, and what limits that spend
• who decides whether the work succeeded

A model can write plausible answers to all of these. It can't commit the business to any of them. A brief filled with answers nobody agreed to is worse than a brief with gaps in it, because when a section is blank, someone has to go back and ask the business stakeholder.

Deciding which customers you won't serve is the clearest example. That choice gives up revenue the business might otherwise have chased. Only the business stakeholder can decide whether that trade is worth making.

The same problem appears with no AI involved. A strategic business brief is where the business stakeholder makes these decisions and writes them down, before marketing plans anything.

Using AI as a Shortcut Pushes the Strategy Work Later

When the strategy isn't settled, many marketers work it out during the creative process, after the brief has already gone to the agency. The agency produces work, the business stakeholder sees it, and only then does anyone say what the campaign was meant to do. A model that produces a finished-looking brief in a minute makes that habit easier to keep.

The cost shows up as rebriefs. An agency that suspects nobody thought hard about a brief will often treat it as a first draft and hold back its best work until the rebrief arrives. Each rebrief adds agency fees and delays the campaign. What a good brief contains, and what agencies say most briefs leave out, is covered in how to brief an agency.

How to Use AI on a Campaign Brief

Get the business decisions first, in writing, from the business stakeholder. No amount of research replaces the stakeholder stating what the work has to return and by when.

Use AI on the evidence once those decisions exist. Summarise the win and loss data, pull the audience research together, and list what competitors claim. Check every figure you plan to use against its original source, because a model can get a number wrong and still sound certain.

Ask AI to find the gaps in your draft. Ask it what a reader couldn't work out from the brief, or which claim has no proof behind it. Finding gaps in a document is a reading task, and models do it well.

Write the proposition and the single objective yourself. These two parts of the brief decide whether the creative team has a clear target or ends up with the category average.

For a wider view of where AI helps marketing work and where people need to make the call, see what AI can actually do in marketing.

What to Decide

Work out where the campaign brief you're about to write is stuck. If you're still gathering evidence, use AI and save the time. If nobody has said what the work must deliver, which customers it's for, or what the business will spend, AI can't help you. Go back to the business stakeholder first.