Thanks to our obsession with using LLMs
to churn out variations on today’s
5X-your-engagement magic prompts,
we’re not just not doing original work —
we’re actively ensuring
BEIGE IS THE NEW BLACK.

When AI first entered marketers’ awareness,
it came with a pretty irresistible pitch:
Look what I can do!
For free!
In seconds!

And if you’re overworked, under-resourced, and just plain beat — 
well, that’s a tough offer to turn down.
Why spend days researching, brainstorming, and iterating
when your shiny new thingy can whip up a decent first draft
faster than you can say “Q2 content calendar”?

So off we went. And for a while, it was fine. Until it wasn’t ... quite.

The problem wasn’t that the original pitch was wrong.
It’s just that it was incomplete — 
and it set the stage for a soft bias of low expectations.

You see, when we only use LLMs to remix the known — 
to parse and repackage what’s already out there — 
we squander what they’re really good at:
spotting patterns, shifting perspectives,
and providing provocations
we might never have reached on our own.

Worse, the very skill that makes LLMs so compelling — 
their fluent mimicry — is becoming their Achilles’ heel.
Especially for marketers.
Because the more we ask them to sound like everything else,
the more they oblige.

And bit by bit, by millions of posts on millions of feeds,
everything starts sounding and looking the same.
Beige. Polished. Predictable.
A pan to a slow fade over a perfectly optimized
Sunset Strip of recycled ideas.

Which brings us to the point -
the real revolution in AI-assisted marketing.
The Holy Grail if you will. It isn’t optimization.
It’s elevation –
not just looking at LLMs as a way to do things faster and cheaper.
But a way to do things better.
That is the answer.

Plus, in a world where most brands are busy automating mediocrity,
there’s real power in being the one who dares to break format,
challenge assumptions, and make something unforgettable.

Let them zig. You zag — and surprise surprise,
your work will actually work.
Not by being faster. Or cheaper.
But through smarts. Ambition. Originality.
And the creative nerve to do what a prompt never can:
Surprise someone.

While a lot of today’s communicators came up in a “if it feels good, do it” era — 
buzz over brand, execution over essence — 
the rise of the LLM has taken me back to what the old-school purists
at places like Ogilvy hammered home.

The fundamentals:
Define the offering and audience.
Translate attributes into benefits.
Align the ultimate benefit with a human truth.
And bring it home with an emotionally-driven execution.

To some, that may sound like common sense.
To others, as outdated as a three-martini lunch.
But in this new world of LLMs, one thing is clear:
The more things change, the more they require us to.

Because while the tools have evolved,
the need for a roadmap is back.
We’ve built a machine that can do the extraordinary
but that machine can only go as far as we humans know how to guide it.

And so while we’re not exactly back in that world
where some guys in Corvettes and flattops radioed,
“Houston, we have a problem,”
we’re somewhere similar.
A place where human beings have to step outside their humanity
and learn to think like a machine.

It’s an approach engineers have long been trained to do - to decompose.
Break a system into parts.
Solve each part in turn.
And now communicators need something similar:
a model for creative decomposition.
One that splits a challenge apart,
isolates the points of friction,
and then identifies and frames the tension worth pushing on.

Put another way, the challenge today isn’t learning how to use AI.
It’s learning how to think in ways AI can elevate.

That’s why we’re proposing something new:
A first principles approach to creativity and strategy — 
reengineered for an LLM-assisted world.
And it starts not with prompts, but with decomposition.

Here's what that would look like in the real world.

Imagine trying to uncover fresh insight into a group — 
say, HR leaders making healthcare decisions.
Instead of asking the model to “generate strategy ideas,”
break the problem apart. Guide it step by step:
map what’s known, probe what’s missing,
surface where expectations clash with reality
and find that one pressure point that makes the strategy matter.

How a strategist could use an LLM to develop insight into HR buyers

STEP	HOW TO USE LLM TO SUPPORT IT	SAMPLE PROMPTS
1. DEFINE THE TARGET AUDIENCE	Capture lived experience, context, and motivations.	“Describe the mindset, daily stressors, and KPIs of mid-level HR professionals at fast-growing startups.”
2. MAP KNOWN ISSUES	Generate a starting list of pain points, then pressure-test.	“List the top 10 challenges HR teams face when selecting healthcare vendors in 2025. Then add 5 counterintuitive ones.”
3. IDENTIFY CONTRADICTIONS OR UNMET NEEDS	Look for where expectations don’t match experience.	“Where might HR decision-makers feel pressure from leadership but lack clarity or control?”
4. LADDER TOWARD INSIGHT	Push on tensions, frictions, and tradeoffs.	“What beliefs or assumptions do HR leaders hold that might be complicating their vendor decisions?”

Same idea but with a different goal.

Say you’re trying to figure out which product benefit truly resonates with a specific audience.
The lazy way?
Ask the model to the write copy.
The strategic way?
Decompose the ask.

Use the model to explore what matters to your audience,
your brand, and your competitors.
Then go further, past the obvious, to not just a solution,
but rather a tension on which you can build a killer campaign.

How creative could use an LLM to develop insight into HR buyers

STEP	HOW TO USE LLM TO SUPPORT IT	SAMPLE PROMPTS
1. FRAME THE PRODUCT ROLE	Capture context for how/why it’s used	“Describe the emotional job this product does for a first-time manager in a remote workplace.”
2. EXPLORE ALL PERCEIVED BENEFITS	List reduce, cluster, and test impact	“List 10 benefits of this product. Group them by emotional vs functional appeal.”
3. SURFACE DIFFERENTIATORS	Identify what’s true and uniquely ownable	“Which of these benefits might competitors also claim? Which are more distinct?”
4. ALIGN TO AUDIENCE NEEDS	Sense-check what resonates most	“For a burned-out HR lead at a Mid-size Utility Co., which point solution matters most? Why?”

While the efficiency gains promised by LLMs
may one day free up budget for creative and strategic talent,
that future’s not here yet.
And there’s little evidence the savings would be reinvested
in better marketing teams anyway.
If anything, the opposite is happening.
Teams are shrinking.
Timelines are compressing.
Leadership expectations are higher than ever.
And the hard truth?

They’re not entirely wrong.
For the vast majority of marketing tasks:
social posts, nurture emails, banner ads, even landing pages,
LLMs often really can deliver something “good enough.”
And so everything seems fine… for now.

But “good enough,” scaled across an industry,
is a slow march toward the middle.
Mediocrity. Beige.
And beige is where brands go to die.

That’s a very real risk. But it’s also an equally real opening.

In a world where sameness is optimized and automated,
difference becomes the secret sauce.
And communicators who know how to work with LLMs,
not just as tools for speed, but as partners in provocation,
could well deliver work so powerfully different
that it will define the next era.

Put another way, the iconic campaigns,
the category redefinitions,
the ideas that make people really feel something
won’t come from AI alone.
But they won’t come from humans alone either.
The future belongs to a new kind of symbiosis — 
between machine intelligence and human judgment.

Which brings us to the real question:
what kind of human is best for that role?

In a world where LLMs can produce a hundred usable ideas in under a minute,
your job isn’t to out-generate them.
It’s to up-guide them so that together you both get to a place
neither of you could have gotten alone.

That starts with upstream thinking.
Not “What do I want it to write?” but
“Where’s the tension worth pushing on?”
The best communicators don’t just prompt. They provoke.
They isolate the rub.
They chat. And challenge.
They think diagonally, leaping between categories,
cultural truths, and emotional drivers.
They connect the unconnected.
And they know how to turn a vague instinct
into a sharp, defensible POV.

They also know when to push.
And when to stop.
When it’s time to throw something out and start again.
When it’s time to test the truth of your LLMs patience.
That’s not about wordsmithing.
That’s about judgment.
And in this new world, judgment is everything.

If you’ve made it this far in the series,
chances are you’ve already got some of what it takes.
Now it’s time to do it — 
not just instinctively, but intentionally.

Fortunately, we’re still relatively early.
Most AI-assisted marketing today looks like early web design:
fast, cheap, and painfully forgettable.
But that won’t last. This is a first mover space.
And the first movers are already moving.

That said, the next wave of great brand work
won’t be fueled by optimization.
It’ll be powered by elevation — 
by thinkers who know how to raise the ceiling,
not just lower the cost.
Who know that ambition, not automation, is what sets work apart.
And that creating such a reality will take purpose and a plan.

So that’s where we’ll go next.

In Article 5, we’ll map the path forward.
From the rituals and roles of high-performing human–AI teams
to getting the help you need in order to put in place the leadership,
the processes, and the mindsets that will ensure you can not just survive
but thrive in this new world.

The future of marketing won’t be automated.
It’ll be authored — by humans and machines,
not just working faster, but reaching higher.

This series wouldn’t exist without the insight, patience, and moral support of two people:

My beautiful wife, Cecile Engrand — the best event marketing CD I know — who was showing the world what was possible with AI long before the rest of us caught on, and whose strategic sensibility still grounds everything I do.

And my lifelong friend, Thomas Bolton — Princeton-trained, fractional CPO, and AI whisperer — who’s been my teacher, tech advisor, and intellectual sparring partner from day one. And the only person I know who’s building his own AlphaGo model … for fun.

Without their very human connection (and the help of my favorite LLM, ChatGPT), none of this would’ve come together.

All film references in this article are used under U.S. Fair Use Guidelines for the purpose of commentary, critique, and cultural analysis. All rights remain with the original copyright holders. If you’re a rights holder and wish to request attribution or removal, please contact me at LiamSherborn@gmail.com.