Why it matters

Your work has to win twice.

Once with the machine that decides what to surface. Once with the human who actually reads it.

I spent twenty years in broadcast marketing and editorial asking one question: will a real person, with no time and no prior commitment, actually get this? That hasn't gone away. It's just grown a second half.

More and more people ask LLMs first. They trust their summaries. They follow up when something cuts through; they move on when it doesn't. LLMs are fast becoming the new search engine. The second half of the question: will the systems that mediate their information get it right first?

Picture the scene.

You publish something new — a report, say. Someone it's genuinely for comes across it. Maybe they read about it in a news article, or heard it being discussed on the radio.

They're busy, it's thirty pages long, and they can't find a decent summary anywhere. So they ask an LLM: “What’s this about?”

Whatever it answers is what they now believe your work says, whether or not they ever open it.

The principle

Build for humans.
Structure for LLMs.

Almost everything that helps a distracted human is the same thing that lets a machine read you accurately. Get the structure right and the same work wins twice. In practice it comes down to three questions.

01 Can a first-time visitor get the point in sixty seconds?

Most arrive meaning to skim, not read. If the core point isn't clear from the headline, the opening and the structure alone, they leave without it — and so does the machine summarising the page.

02 Does their attention compound?

Attention compounds when each piece carries the visitor to the next — so someone who came for one thing finds their way to the rest of what you've made, and their interest builds instead of resetting to zero. Most sites win the attention, then waste it.

03 Would a machine reading it cold land on the same meaning a careful human would?

A machine can only pass on what it can read. Structure your work clearly and it relays your argument intact; bury the point and it relays the confusion.

A common mistake.

Organisations — especially those looking to shape public opinion — will often publish their flagship report as a PDF and link to it from the site. It feels like the natural home for serious work. But a link to a PDF is a closed door. The hurried human weighs up whether to download thirty pages and often doesn't bother; the machine summarising the page may not open it either. The argument ends up behind the door instead of in front of the reader. The same report, surfaced in well-built HTML, walks straight through.

A recent experiment

In April 2026, Freedom in the Arts published its flagship report, The New Boycott Crisis, as a PDF linked from their site. A speculative Red Slash rebuild put the same argument on the page, in structured HTML. In June 2026 I asked the three tools people now research with — ChatGPT, Claude and Perplexity — to summarise the report from each surface.

From the rebuilt page, every tool gave the current report back accurately, first time, straight off the page. From the original, it was a lottery: about half the time — and even the same tool on different runs — what came back wasn't the flagship at all, but an older report whose text sat on the page, while the current argument stayed unread inside the PDF. A capable machine can read a PDF; the catch is you're relying on it to choose to.

What’s at stake

Leave your argument behind a link, and you leave your reach — and your narrative — to chance.

Put it where both readers land — on the page, built to be skimmed and built to be read — and nothing rides on whether a machine decides to open a download.

Where I come in.

This doesn't replace how you already work — it adds a new front: making the same work hold up on a route to your audience that barely existed a few years ago. It starts with a cold read of your current site, and a clear set of recommendations. From there it's your call: take them in-house, or have me carry them through.

It's the job I've always done — making sure good work reaches the people it's for. And it's one piece of something larger: most of what I do is help people make more from what they've already made, and winning twice is how that work reaches people now.

Is your work getting through?

Tell me what you've made. I'll show you what an AI makes of it today — and what the people you're trying to reach are missing.

hello@redslashstudio.com

or