The homepage is no longer the front door. The question is whether you have a wormhole or just a wall.

How AI traverses

When AI evaluates a destination on someone's behalf, it does three things in sequence:

1. Identity compression. Builds a quick mental model of who you are, what you do, and who you serve. If the model is ambiguous, AI stops here. You don't make it past first contact.

2. Confidence. Decides whether your signal is strong enough to trust. Has the claim been tested? Is the expertise specific? Are the supporting details consistent across the site? Weak signal here, AI doesn't go deeper.

3. Expansion. If compression and confidence both succeed, AI explores the supporting depth — the intent magnets, the long-form pieces, the proof. This is where it builds the recommendation it'll bring back to the user.

Most sites die at step 1.

What strong signal looks like

A site optimized for the new physics has:

  • Clear category ownership. One sentence on what you actually do, repeated in slightly varied form across pages.
  • Pillar-and-subtopic clustering. A central manifesto with supporting depth in concentric rings.
  • Mechanism plus evidence. Not just claims, but how it works and why anyone should believe you.
  • Consistent terminology. The same things called the same names across the site.

A site killing its own signal has:

  • Generic "we help X with Y" copy.
  • Unconnected articles published on a chronological feed.
  • Different terminology for the same thing in different places.
  • Benefits floating in the air with no mechanism behind them.

The difference is recognizable in seconds — by AI and by humans.

The five traps

Common patterns that look like content strategy but actively damage AI legibility:

Chronological blog feed. A list of dated posts in reverse order. Tells AI your site is a stream, not a body of work. Hard to read as expertise.

Cross-posted social content. Twitter threads, LinkedIn posts, recycled newsletter snippets. Reads as noise. Dilutes the signal of your actual claims.

Generic "Insights" hub. Articles about industry trends with no clear point of view. Hard to compress into an identity.

Market commentary. Reactions to the news of the week. Same problem — no through-line, no claim.

Full newsletter archive. A linear log of past sends. Useful internally; bad as a public surface.

If you're doing any of these, you're publishing for an audience that no longer exists.

The book beats the blog

A book is organized by topic and stance, not publish date. It has a clear argument. Each chapter supports the argument. The whole works because the parts work together.

A blog is organized by recency. The connective tissue is the publish schedule, not the ideas.

In an AI-mediated world, the book wins. AI can read it, compress it, recommend it. The blog gets crawled for keywords nobody is searching for anymore.

The implication: structure your site as a book. A sharp manifesto. Supporting depth that reinforces it. Specific intent magnets for the highest-leverage questions your audience brings. No chronological feed unless you genuinely have a publication pace and a reason for it.