Designing for Humans and Machines: The Two Internets We’re Building Now

The Internet Has a New Reader

AI doesn’t browse your site the way a human does. It doesn’t admire your design, infer your intent from visual hierarchy, or “get the gist” from a compelling hero image. It sends an agent — a program that scans structure, parses hierarchy, and looks for explicit signals. Then it makes a fast decision: Can I confidently explain what this organization does?

If the answer is unclear, even if the site is beautiful, the agent moves on. This shift is still early, but it’s accelerating fast. The organizations getting ahead of it now are the ones building for both readers simultaneously.

Crawlability Is the New Baseline

AI agents already navigate websites, compare offerings, read content, and complete actions on behalf of users. That’s not a future state. It’s happening now in tools like Perplexity, ChatGPT, and emerging agentic frameworks, like Claude CoWork, which can research vendors, draft RFPs, or shortlist providers without a human doing the clicking.

That changes the bar for what a website needs to do. A site that’s passive — beautiful to look at, designed for human browsing — can fail completely when an agent tries to extract structured meaning from it. Websites now need to be machine readable, navigable, and actionable for the AI acting on behalf of people, not just for the people themselves.

Accessibility Is the Shortcut

Here’s the fastest diagnostic available: run an accessibility audit on your site.

The same principles that make a site accessible to users with disabilities — clear semantic HTML, logical heading hierarchy, descriptive link text, and labels that name interactive elements — are precisely what make a site legible to AI agents. Both need explicit structure. Both struggle with ambiguous navigation. Both fail when meaning is conveyed through visual design alone.

An accessibility score is increasingly a proxy for AI readiness because they’re testing the same underlying thing: whether meaning is encoded in the structure of the page, not just in how it looks. Organizations that have invested in accessibility over the years have, often without realizing it, built a significant head start on GEO readiness.

The Three Questions AI Asks

Every AI agent — whether it’s assembling a search answer, powering an AI assistant, or executing a task on behalf of a user — is effectively asking three questions of your site:

1. Can I understand this?
   Is the structure clear? Do headings mean something? Is the content organized in a way that signals what this organization does?
2. Can I find what I need?
   Is the information architecture logical and shallow enough to navigate without inference?
3. Can I act on this?
   Do forms and CTAs clearly describe outcomes? Can an agent understand what happens next?

If any answer is “maybe,” the agent often bails — and takes the user’s attention with it.

This Isn’t Anti-Design. It’s Pro-Clarity.

There’s a common misconception that designing for machine legibility means stripping out the things that make digital experiences feel human — narrative, emotion, brand presence. It doesn’t.

Humans still need emotional connection. Brands still need presence and the ability to signal quality. But clarity is now a design constraint, not a technical afterthought. The structure of a page, the hierarchy of content, the explicitness of navigation — these are decisions that affect both audiences simultaneously.

The best digital experiences do two things at once: they inspire humans and instruct machines. That’s not a compromise. It’s a higher bar.

GEO Brings Infrastructure Back into Focus

Structured data, semantic markup, and clarity of intent aren’t invisible plumbing anymore. They are — directly and measurably — what determines whether your brand shows up when AI systems assemble answers.

SEO was about being found. GEO is about being understood. And the infrastructure that supports understanding — semantic HTML, structured schema, clear hierarchy, explicit content signals — is now the most strategically important layer of your digital presence.

Get help making your site understandable to AI — not just impressive to humans.

One North works with organizations to identify where AI agents get lost, blocked, or confused, and to apply Generative Engine Optimization strategies that improve clarity, accessibility, and machine readability without sacrificing brand experience.

Photo Credit: Ian Tuck | Unsplash