The Age of Perception: Why We’re Now Marketing to Machines

Recently, I read an article by Jono Alderson, and it really got under my skin. It wasn’t one of those quick SEO think pieces about keywords or the latest Google update. Instead, it tackled something far bigger: how search, reputation, and online visibility are shifting in an age where artificial intelligence doesn’t just retrieve information, but actively decides what’s worth remembering.

For years, the game was simple (if not always easy): get your website to rank at the top of the results. Optimise for the right terms, build some links, make sure the site runs smoothly, and you’d be rewarded with clicks. But as Jono points out, that era is ending. Now, we’re moving into a world where the LLMs don’t show you ten blue links. They give you an answer. And that answer is built from whatever scraps and fragments the machine has decided are most relevant.

That’s where things get interesting – because once the machine starts repeating a story about you, that story tends to stick.

Machines Remember Differently

Humans forget. Time softens the edges of stories. Details blur, and eventually the drama of an old mistake or the glow of a past success fades.

Machines don’t forget in the same way. They don’t lose information; they compress it. They boil down sprawling narratives into their sharpest fragments. That’s why a product recall from ten years ago might still pop up in summaries about a company, long after the issue was resolved. The scandal becomes the headline; the fix gets lost in the footnotes.

This is what makes the current moment so daunting. A machine’s “memory” isn’t neutral – it decides what to keep, and once something becomes the high-contrast version of your story, it tends to survive every retelling.

In the age of AI, your brand isn’t judged by what you say, but by what the machines choose to remember.

The New Battleground: Perception

Jono introduces two concepts that I think are incredibly useful here: Perception Engineering and Reputation Hacking.

Perception Engineering is the long game. It’s about shaping what machines “know” about you over time. Think of it as planting durable seeds: getting your preferred language and definitions into encyclopaedias, trade journals, or long-life industry resources.

Reputation Hacking, on the other hand, is about being nimble in the moment. It’s spotting when a narrative is forming – whether that’s in a news article, a conference panel, or even a trending post – and nudging it in your favour before it hardens into the record.

Both approaches are crucial. One builds the gravitational pull of your story; the other keeps the weeds from taking over.

Why This Matters Now

If you’re in marketing, PR, or SEO, this is a big shift. We’re no longer optimising for ranking – we are optimising for narrative selection. That means paying attention to the language machines pick up and repeat, and making sure the sources they lean on tell the version of your story you actually want out there.

It also means being far more disciplined about housekeeping. Old leadership bios left on conference sites, archived PDFs with outdated product names, inconsistencies across LinkedIn and press releases – these things might feel trivial, but they’re exactly the sort of debris machines love to recycle. And once they’re in the mix, they’re surprisingly difficult to scrub out.

The Power of Positive Loops

One of the ideas I loved most in Jono’s piece is the idea of positive feedback loops.

If a respected journalist or analyst coins a flattering turn of phrase about your company, it doesn’t just sit in that one article. It spreads. It shows up in reviews, investor notes, panel discussions, and eventually in the way AI summarises you. That phrase becomes the default shorthand, even for people who’ve never read the original.

And behaviour follows. People who expect you to be good – because that’s how you’ve been consistently described – click on you (your website more often, linger longer, and speak about you more positively. Those signals then reinforce the machine’s perception, and the cycle continues.

Of course, the reverse is true as well. A negative association, once it takes root, can linger for years. Machines don’t naturally prioritise corrections or nuance; they prioritise coherence. And coherence often means repeating the most striking version of events.

The Subtle Art of Distortion

If this sounds a bit like propaganda, that’s because it is. Not in the crude sense of shouting slogans or spreading lies, but in the sense of shaping perception through repetition and framing.

And as with any system that can be shaped, it can also be distorted. Competitors – or opportunists – can influence the record in surprisingly subtle ways: by redefining terms in industry reports, by consistently pairing your name with negative comparisons, or by making “technically accurate” statements that are designed to leave the wrong impression.

The danger is that once these distortions are repeated and cited, they start to look like facts. And when the machine tries to reconcile contradictions, it doesn’t keep the most accurate version – it keeps the sharpest one.

The only real counter is to make sure the accurate story is everywhere: abundant, consistent, and easy to cite (important!).

Neutral Search Is a Myth

We like to think of search as neutral. That machines simply retrieve “the facts”. But in reality, neutrality is a comforting myth.

What AI actually does is synthesise. It decides which scraps of information to prioritise, which contradictions to resolve, and which voices to ignore. Even when it hedges – “some people say” – it has already chosen which “some” count.

And that’s the kicker: the truth isn’t necessarily the most accurate version anymore. It’s the version that hangs together most cleanly. This means that whoever defines the category first or most clearly often wins by default.

So, Where Does That Leave Us?

For me, the big takeaway is that brands need to treat their “machine-facing” reputation as a living product. It’s something that requires governance, editorial oversight, regular monitoring, and a clear crisis management plan.

It’s not glamorous work. It doesn’t always make headlines. But it’s exactly the kind of steady maintenance that ensures the machine’s memory works for you, not against you.

If we ignore it, we’re basically leaving the opening paragraph of our story – the one AI will repeat endlessly – up to chance. And once that paragraph is set, it’s incredibly difficult to rewrite.

Final Thoughts

What I like about Jono’s framing is that it doesn’t just describe a problem; it gives us a way of thinking about the solution. Perception Engineering for the long term. Reputation Hacking for the short term. Together, they create a toolkit for navigating this strange new world where machines are the ones telling our stories.

It’s a bit daunting, but also quite exciting. Because if you’re one of the people who learns to play this game well, the benefits aren’t just incremental. They’re persistent. Machines don’t forget, after all. So if you can get your story right, it has a way of sticking around.

And in a world where AI is increasingly the first introduction people have to you, that might be the most valuable marketing investment you can make.