Being exceptional in private is no longer enough

The report wasn't subtle. Missing context. Thin content. Pages that were beautifully designed but gave AI nothing specific to hold onto. A hero section that described only a feeling rather than a capability too. Service pages that could have applied to any agency in any country.

AI doesn't experience your website the way a visitor does. It doesn't notice the photography, feel the weight of the typography, or register that your brand feels premium. It reads the underlying text, the headings, the content, the structure and it asks four questions:

• Have you actually done this?
• Do you know your domain deeply?
• Do other credible sources reference you?
• Are you consistent and honest?

Google calls this E-E-A-T: experience, expertise, authority, trust. AI has inherited that framework entirely. And if you can't answer yes to those questions in a way that's publicly visible on your site, AI can't recommend you. It doesn't matter how good you actually are.

The most interesting problem in AI-era web design isn't technical. It's the same tension that's always existed between designers and developers. Designers push visual boundaries. Developers apply logic and best practice. Neither wins completely, and the best outcome lives in the negotiated middle.

AI readability creates the same push and pull. A client might love a full-screen visual experience with minimal text and from a brand perspective, that restraint can be exactly right. But if AI reads that page and finds nothing specific, the business doesn't get recommended. The sweet spot is a site that's visually compelling and structurally legible. That's harder than doing either one alone.

It's not about making sites look worse. It's about making sure the words and structure underneath the design are doing real work. A beautifully designed page with specific, credible, well-structured content performs better for both visitors and AI. The constraint makes the design better, not worse because it forces clarity.

"Sustainable packaging" means nothing to AI. "Compostable primary packaging certified for direct food contact" is a citation waiting to happen. "Luxury experience" means nothing. "An intimate twelve-guest property with a chef-led dining programme sourcing exclusively from the local region" that's something AI can work with.

Vague claims don't get cited. Generic positioning doesn't get recommended. The businesses that win in AI search aren't necessarily the biggest or loudest they're the ones with the most clarity about who they are, what they do, and who they do it for. That clarity has to live on the page as real text, not locked in a PDF or a sales conversation or a rep's head.

This is partly a brand problem. Strong positioning has always mattered but now the stakes are higher. If your positioning is vague, AI can't build a confident picture of you. And if AI can't build a confident picture, you don't make the shortlist.

Your enquiries might still be coming in. Things might feel stable. But what you can't see is the shortlist you're not on. The buyer who asked AI, got three names back, and yours wasn't one of them. The decision made before anyone picked up the phone. Before they attended your trade stand. Before your sales team even knew the opportunity existed.

Stability in your current numbers isn't evidence that nothing has changed. It might just mean you haven't felt it yet. Authority compounds, citations build on citations, reputation reinforces reputation, and the businesses building this now will establish a lead that's genuinely hard to close later.

If your site is already due for a rebuild, AI readiness is one of the strongest reasons to move now. The architecture decisions you make at the start, framework, CMS, content structure , determine how readable your site is to AI for the next several years.