One question I ask AI before every brief

Something shifted in how I write briefs.

Before I lock anything in, I paste the draft into Claude and ask one question: what assumptions am I making here that I haven’t actually proven yet?

What comes back is usually the part I didn’t want to look at. You’re assuming the audience already understands what this product does. You’re assuming the problem is awareness when it might be trust. You’re assuming this needs to be on Instagram because that’s where you’ve always done it.

None of those are things a brief template catches. They are the things a good strategic partner would push back on in a meeting, except that meeting rarely happens before production is already moving.

I started doing this after a campaign where we built something genuinely strong creatively, and it still didn’t work. The brief was never challenged. Everyone liked the direction, so we moved. The wrong assumption was baked in from day one, and nobody caught it because nobody was asked to look for it.

Now I treat that 20-minute conversation with an AI as the cheapest possible way to find the hole before it becomes expensive. It doesn’t replace thinking. It means the thinking gets pressure-tested before it hardens into a production schedule.

The brief is where most campaigns are won or lost. We just don't treat it that way.

What part of your process never gets properly challenged before it’s too late to change?

Originally published: LinkedIn, April 2026