6 min read · 1,327 words

Eighteen months into using AI tools inside a brand function, the honest inventory looks nothing like the vendor pitch decks. Some things genuinely changed — first-draft speed, research synthesis, localisation across markets. Most of what actually matters in brand work — judgement, category insight, knowing which stat to cut because it’s not quite right — didn’t move at all.

I started using generative AI tools seriously for brand and communication work in 2024, first for first drafts, then for competitive research synthesis, then reluctantly for creative concepting because everyone on the team was already doing it quietly. The gap between what I expected and what actually happened is the most useful thing I can report back, because most of what’s written about “AI in marketing” right now is either breathless or dismissive, and the reality sits in a much less exciting middle.

What genuinely got faster

First-draft speed on long-form content is the clearest win, and it isn’t close. A brief that used to take a writer three to four hours to turn into a rough draft now takes 30 to 45 minutes of AI-assisted drafting followed by an hour of substantive editing — the editing hasn’t gotten shorter, but the blank-page time has nearly disappeared. Competitive research synthesis is the second real win: pulling together what ten competitors said in their last four quarters of public communication used to be a day of manual reading; it’s now a couple of hours of directed research plus verification.

McKinsey’s 2024 research on generative AI in marketing functions found meaningful productivity gains concentrated specifically in content generation and summarisation tasks, which matches what I’ve seen directly — the gains are real, but they cluster tightly around a specific set of tasks rather than spreading evenly across the function.

What didn’t change, no matter how good the tool got

Positioning didn’t get easier. Deciding what a brand should actually stand for, in a specific category, against specific competitors, for a specific buyer who has specific frustrations — that’s still slow, still requires sitting with customers and sales teams, and AI tools are actively bad at it because they default to generic, defensible-sounding positioning that could apply to almost any company in the category. I’ve had to explicitly fight a tool’s own instinct toward blandness more times than I can count.

Category judgement is the second thing that didn’t move. Knowing that a stat is technically accurate but will land wrong with a specific board, or that a campaign concept that tested well will read as tone-deaf in a market that just went through layoffs — that’s pattern recognition built from years of being in rooms, and no tool I’ve used gets close to it. Gartner’s research on generative AI in marketing organisations has flagged this specific gap — the tools accelerate execution but don’t substitute for strategic judgement, which sounds obvious written down and is routinely ignored in practice by teams under budget pressure.

The AI slop problem is real, and it’s a brand risk, not just a quality complaint

The visible failure mode across the industry right now is generic, interchangeable content — every company’s LinkedIn post reading like it came from the same three prompts, because it largely did. The Content Marketing Institute’s research on AI-assisted content has flagged the same pattern industry-wide: audiences are getting faster at pattern-matching AI-typical phrasing and disengaging from it within a sentence or two. HubSpot’s own State of Marketing research has found a similar split — marketers report AI saves time on production, but differentiation and originality remain the top-cited concerns about over-relying on it. For a brand function specifically, that’s not a productivity problem, it’s a differentiation problem: if your content sounds like everyone else’s, AI hasn’t helped your brand, it’s actively eroded the one thing brand work is supposed to build.

Where this leaves brand teams in practice

The teams doing this well aren’t the ones using AI the most — they’re the ones using it for a narrowly scoped set of tasks and holding the line hard everywhere else. Draft acceleration, research synthesis, localisation across markets where a human reviewer still signs off — yes. Original positioning, brand strategy, anything that needs to sound distinctly like a specific company rather than a category-generic version of it — no, or at minimum, AI-assisted-then-heavily-rewritten rather than AI-generated. I wrote about the underlying visibility mechanics that make distinctive, fact-dense content more valuable now, not less, in the future of marketing and navigating the digital landscape.

Does this mean brand teams need fewer people?

Not in my experience so far — it’s changed what the team’s time goes toward, not the size of the team. The hours saved on first drafts and research synthesis have mostly been redirected into more original strategic work and more actual customer conversation, which was already the scarcest resource on most brand teams before any of this started. Headcount pressure in marketing right now is a budget story more than an AI-substitution story — LinkedIn’s Economic Graph data on marketing job postings tracks that pressure moving in line with broader marketing budget cycles, not with AI adoption curves specifically.

What this means for you

Audit your own team’s AI usage against the same split: draft speed and research synthesis are safe to hand over broadly; positioning, judgement calls, and anything meant to sound distinctly like your brand should stay firmly human, AI-assisted at most. The teams that blur this line are the ones producing the interchangeable content that’s starting to actively hurt brand differentiation rather than help it — a risk I think is under-discussed relative to how much airtime “AI productivity” gets. I’ve made a related argument about protecting what’s genuinely differentiated in data-driven marketing and unlocking insights.

Frequently asked questions

Has AI actually made brand marketing teams more productive?

Yes, specifically for first-draft content generation and competitive research synthesis, where blank-page and reading time drop substantially. It has not meaningfully accelerated positioning strategy or original creative judgement, which remain the bottleneck tasks.

What is “AI slop” and why does it matter for brand?

It’s the industry term for generic, interchangeable AI-generated content that readers increasingly pattern-match and disengage from. For a brand function, it’s a differentiation risk, not just a quality issue — if your content sounds like every competitor’s, AI has quietly worked against the brand’s core job.

Should brand teams stop using AI tools for creative work?

Not stop — but treat AI output as a starting point for ideation volume, never a finished creative concept. The judgement on which idea genuinely fits the brand and moment still needs to be human, and skipping that step is exactly what produces generic output.

Has AI reduced the need for brand marketing headcount?

In my experience, no — it’s redirected time saved on drafting and research toward more original strategic work and customer conversation, rather than shrinking the team itself. Headcount pressure in marketing right now tracks budget cycles more than AI substitution.

What’s the biggest mistake brand teams make with AI tools right now?

Using them for positioning and strategic direction rather than execution support. AI tools default toward generic, category-safe answers, which is close to the opposite of what distinctive positioning requires.

Where has AI actually changed your team’s day-to-day work, versus where does it just sound like it should have? I’d like to compare notes with other marketers who’ve been in it for more than a quarter.