AI Won't Fix a Vague Brief

There’s a new tool announced somewhere every week. Most of them are variations on the same pitch: generate content faster, score your leads automatically, write your emails while you sleep. The packaging changes. The underlying promise doesn’t.

Some of it is genuinely worth paying attention to. AI-assisted drafting has become a real part of how I work with clients now. For first-pass copy on email sequences, landing pages, or research summaries, it cuts down time in a way that actually matters for smaller teams. If you’re a three-person marketing function trying to do the work of six, tools that help you move faster without adding headcount are worth taking seriously.

But there’s an assumption baked into a lot of the enthusiasm that I think is worth examining.

Speed Isn’t Usually the Real Constraint

The businesses I see struggling with their marketing are mostly struggling not because they’re slow, but because nobody’s agreed what they’re trying to achieve, or who they’re actually trying to reach. The brief is vague. The audience is loosely defined. The message is somewhere between "we do lots of things" and "we’re quite good at them."

Feeding that into an AI tool doesn’t produce clarity. It produces polished vagueness at scale, and that’s already what most of their competitors are putting out. The content looks fine. It reads fine. It doesn’t land with anyone because it wasn’t really aimed at anyone.

I’ve seen this with a client in professional services who adopted several AI content tools in fairly quick succession. Output went up. Enquiries didn’t. When we sat down and worked through who they were actually writing for and what those people cared about, the tools became genuinely useful. The same technology, pointed at a clearer target, started producing things worth sending.

The tools hadn’t changed. The thinking behind how they were being used had.

What AI Is Actually Good At

It’s worth being specific about this, because the category gets talked about as if it’s one thing.

AI is good at producing. It can take a direction and execute against it quickly, iterate on tone, restructure an argument, generate variations. For someone who knows what they want to say and who they’re saying it to, that’s a real advantage. The time between "here’s the thinking" and "here’s a decent draft" has compressed considerably, and that matters.

What it isn’t good at is deciding what’s worth producing in the first place. It can’t tell you whether your positioning is differentiated or whether it sounds like everyone else in your sector. It can’t judge whether the message would actually resonate with the person who holds the budget sign-off, because that judgment requires understanding context, history, and the specific way that person thinks about their own situation. It requires experience of what tends to work and what tends to get ignored, and some instinct for the difference between those two things.

The businesses I see getting real use out of these tools are the ones who had clear enough direction that the tools had something useful to execute against, not the ones who adopted everything earliest. That clarity didn’t come from the tools. It came from doing the harder, slower work first.

The Thinking Still Has to Come From Somewhere

Every quarter the tools get faster and more capable. That’s genuinely true, and it’s worth keeping up with. But the thinking that makes them useful hasn’t been automated, and I’m not sure it will be any time soon.

Deciding what a business should be known for, which conversations it should be starting, which customers it’s actually best placed to serve, that’s still judgment work. It’s the kind of thing that takes time and usually involves some uncomfortable honesty about what’s not working. No tool speeds that part up, because it’s a thinking constraint, not really a production one.

The businesses that will get the most out of AI over the next few years are probably the ones who recognise that distinction now, and don’t mistake faster output for better direction.

---

Common questions

We're already using AI tools for content and it's not making much difference. Where do we start?

Usually the starting point is the brief, not the tool. It's worth stepping back and being specific about who you're writing for and what you want them to think or do differently after reading it. Once that's clear, the tools tend to become a lot more useful.

Is AI-assisted content drafting something smaller B2B teams can realistically use, or does it need a specialist to manage it?

Most smaller teams can use it without needing a specialist, but it works better with some structure around it. Having a clear sense of your audience, your tone, and what a good piece of content actually looks like for your business means you can direct the tools rather than just accept whatever they produce.