Your Brief Sounds Complicated. It's Saying Very Little.

“Briefing… AI or not to AI” : a blog for strategic, internal/HR, SME, marketing comms teams and anyone putting briefing docs together really…

The brief used to take two hours and force you to confront what you didn't know. Now it takes five minutes and confronts you with nothing.

Before AI, an unfinished brief looked unfinished. Either it was vague or it had picked a direction with no reason behind it. Both told us, in about 30 seconds, that the real conversation hadn't happened yet.

A year ago you could spot an AI-written brief quickly. Buzzwords stacked like pancakes. Em dash parties against negative parallelisms. Every other sentence ending on "transformational." And a paragraph somewhere helpfully explains that a video is "imagery on the screen set to music."

Cool, thank you, I'll be sure to add imagery to the screen.

False certainty is the real problem here, more than AI itself. And that problem has evolved. The briefs landing in our inbox today read better than they did 12 months ago, with a tighter lexicon, smoother rhythm, and the obvious tells mostly sanded off. Underneath, the thinking is still hollow. The briefs that arrive over-confident still tend to be under-thought.

We see this most in corporate and government comms, where strategic teams have folded AI into their workflow faster than almost anyone else. The briefs we receive arrive structured and with the confidence of a salesman. The clarity is a costume, and it tends to do the damage later in the project.

Now, none of this is a failing on your part.

You're a small business that's been told AI will triple your skill set this year, and now you're time poor. Or you're inside a large organisation working out the best way to land a piece of engagement that's been done a thousand times before, and you need to work it out.

We get it, and we want to navigate this new landscape too.

The new tell is sophistication

AI used to pad with corporate mush. Now it pads with what reads like considered prose, full of flowery language, sentence variety, and a bit of rhythm. The fluff has texture, which makes it harder to clock at a glance and easier for everyone in the approval chain to nod along to.

The words have shifted too. The old kit still applies (transformational, future-focused, journey, stakeholders, innovation), but the newer models are smart enough to swap the most obvious ones for something more elegant. So now we get briefs talking about "compelling narratives that resonate," "human-centred storytelling that meets audiences where they are," and "purpose-driven messaging anchored in lived experience." Paragraphs with zero specifics, written in the voice of the guy who narrates every 90s film trailer the same way.

IN A WORLD WHERE EVERY VIDEO IS PROFOUND… YOUR VIDEO CAN BE TOO.

The AI follows it up with generic video rules. Hook in the first 5 seconds. Keep it under 90 seconds. Strong soundtrack. Subjects speaking directly to camera. All of which are rules that apply to some type of video. The question is whether they apply to your video, and AI can't answer that without a clear purpose feeding it. Without the purpose, the rules are decoration, and the brief lands looking like it's been written by someone who knows what they're doing.

This is the hard part of any brief: having a point of view that's your own, the angle on your organisation that makes one piece of work distinguishable from the next. AI doesn't carry that for you. Substance is your job. If you outsource that part to AI, the brief comes back looking finished while the project pays for it later.

The brief feels done. The thinking isn't.

This is the sneaky one.

A brief arrives with full narration, a visual plan, taglines, a three-act structure, illustrated mockups, all from AI.

Plus the client has been sold on a concept they only just encountered, because "look, it's so pretty and the AI says this is a great approach."

Our job at that point becomes "can you produce this, but better?", which is how you end up rewriting while production is already in motion.

It's painful for everyone. The client feels like the creative team is picking at something that's been approved. The creative team feels trapped polishing a direction that was never properly chosen.

When you ask AI "what do I need for a video?" or "here's our idea, what would you do?", it answers confidently and completely. AI gives you a defensible average. That's useful for some jobs but lethal for the one job a brief actually has, which is to be specific to you.

Last year we had to renegotiate scope mid-production because the language baked into a VO script took weeks to unpack. Once we did, we could see why the messages weren't landing. The brief had decided what the video was saying before anyone had asked whether that was what the audience needed to hear, and by the time we got to the truth the video was already in the studio. That's the cost of a brief that arrived too polished to question.

More pages doesn't equal more clarity

The same problem hides at higher prep levels too, just dressed up better.

We've seen strategy decks loaded with content that doesn't help us make creative decisions, because the insights are either obvious or impossible to action. We've seen storyboard docs that map every scene and caption before anyone's pressure-tested the central idea.

One example. A brief came in with the format pre-decided: interview-led, subjects speaking to a key-message-heavy script, every line accounted for. The script was tight enough that the subjects couldn't be themselves on camera, and without that you don't get the moment of real that lets an audience in. The format had been chosen before the question of what would actually move the audience was asked.

The goal of a brief is to clear the path for the best idea to arrive. Often the best brief is half the length of the worst one, with twice the specificity.

What a good brief actually does

Before AI got involved, a good brief was rough and half-formed on ideas, but packed with details. Or better yet, questions. And it was honest.

A good brief gives ingredients a creative team can cook with, and leaves room for the conversation that turns those ingredients into a recipe.

The best ones we receive are clear on the objective, honest about the real internal problem, specific about the audience, transparent about constraints, and strong on references (the work the client loves and the work they hate, with reasons attached). They're willing to admit when a direction is still up for grabs.

How to actually use AI on a brief

AI is genuinely useful on the client side when you point it at the right job. It pulls messy notes into a one-pager. It spots gaps you've left in your own thinking ("you haven't said who this is for"). It structures chaos into something legible, and it asks you the preliminary questions a producer would ask on the first call.

A useful prompt when you're at your desk trying to figure out how to brief us:

You are a creative producer who has made 200 brand films. Interview me, one question at a time, until you have:
- the single behaviour change we want from the audience
- the internal political reason this video is being made
- three references I love and three I hate, with reasons
- the constraint I'm afraid to tell my agency.
End by writing the brief in 250 words, then list the three questions I dodged.

The job becomes pulling clarity out of you, so we can have a real conversation the minute the meeting starts.

The one-page checklist that saves everyone time

If you can send a creative team a single page covering these, you've done your job. Print it, fill it in, send it. (And not like this guy.)

And yes, AI will fill this template for you. The job of the template is to force the questions, not to keep AI out of the room. But make sure you're really thinking about each suggestion and you're feeding it enough thought out information.

• Audience. Be specific about which staff/customer, what they already believe, and what they're sceptical about.
• The single action. What's the one thing you want them to do after watching?
• The real internal/external problem. What's actually happening inside the organisation that makes this video necessary? What issue does the product face?
• Three proof points. What makes the message credible?
• Tone. Three adjectives for the feeling you want, plus three "avoid this" adjectives for what we should steer clear of.
• Non-negotiables. Legal, brand, must-includes.
• What you're bored of and what you hate. Be specific so we can avoid serving it back to you.
• References you love and references you hate. A sentence on why for each.
• Logistics. Key dates, approvals, distribution channels, accessibility requirements.
• Budget range. Even a loose band helps us come back to you with options.

That's enough to start a proper conversation, and none of it needs AI to dress it up for you.

Where we're landing

We're not sure we've cracked this. We're noticing it, talking about it internally, and writing this to find out what other people are seeing.

Again, we know false certainty is an actual problem here, more than AI itself. AI that helps you get clearer is gold. AI that does your thinking for you, or makes your brief look finished before the thinking is finished, is where good projects quietly go mediocre.

It's about the fundamentals of what you want for a video, and making sure you are listing all the things that might be of interest to why you want to make a video in the first place. Not the other way around.

Over-worded, over-confident briefs cost the project time. We end up reverse-engineering an interesting version of the obvious one, which is the slowest and most expensive way to land on good work.

We're not asking anyone to send us less. Bring us polished briefs, bring us rough ones, bring us ideas. AI ideas count too. We'll work with whatever you've got. What we'd ask is that you're clear and honest about which parts are yours and which parts the AI added, and that the decisions you've locked in early are still open enough to talk through with us.

AI is changing what we get paid for, and that's something we're still working out too. It seems to be pushing us toward the part of the work AI can't do: taste, judgement, the room to disagree with the obvious answer. That part, we can defend. The rest, we're learning as we go, same as you.

So, a question. If you've briefed a creative team for anything (corporate, brand, internal, government, NFP), we'd love to hear what's been hard about the briefing process for you lately, and whether AI has changed it for better or worse.