AI Discoverability Is an Information Architecture Problem

Key Takeaways

• AI visibility starts with understanding. Before optimizing for citations or mentions, make sure machines can clearly understand who you are, what you do and where your expertise sits.
• Think beyond individual pages. Products, problems, use cases, industries, expertise and evidence should form a connected information ecosystem.
• More content doesn't automatically create more authority. Consolidating duplicated or fragmented information can sometimes be more valuable than publishing something new.
• Internal links communicate relationships. They should connect genuinely related expertise, solutions, evidence and outcomes rather than being added simply for SEO.
• Schema reinforces clarity rather than creating it. Structured data works best when the underlying architecture and content already make sense.
• SEO should influence information architecture early. Search intent, buyer language and content relationships should help shape the website before development.
• Modern websites need to work for people and machines. A strong architecture makes the organization easier for buyers, search engines and AI systems to understand.

Why better AI visibility starts with making your business easier to understand

There is a tendency in AI search right now to look for the optimization tactic.

Add an llms.txt file. Rewrite content for answer engines. Add more schema. Publish more FAQs. Track citations in ChatGPT. Create pages targeting prompts rather than keywords.

Some of those tactics can be useful, but I think they risk distracting us from a much more fundamental issue.

If an AI system cannot clearly understand your business, its services, expertise and relationship to the problems people are researching, why should we expect it to surface your brand consistently?

This is why I increasingly believe AI discoverability is partly an information architecture problem.

For years, information architecture has been treated primarily as a UX and SEO consideration. We organize pages into logical hierarchies, create taxonomies, establish internal linking relationships and make content easier for users and search engines to navigate.

In an AI-mediated discovery environment, that structure becomes even more important because we're not simply trying to get a page indexed. We're helping machines develop a coherent understanding of the organization behind those pages.

AI systems need to understand more than individual pages

Traditional SEO has often encouraged us to think at the page level.

Which keyword does this page target? Is the title tag optimized? Does it have sufficient content? What internal links point toward it?

Those questions still matter, but AI discovery introduces another question:

What does this page tell a machine about the wider business?

Imagine a cybersecurity company with separate pages covering managed security, MDR, endpoint protection, threat detection and incident response.

If those pages exist independently, use inconsistent terminology and provide little indication of how the services relate to each other, the website may contain plenty of relevant information while still presenting an unclear picture of the company.

Compare that with a structured ecosystem where the relationships are obvious:

That structure is useful for a buyer navigating the website, but it also gives search engines and AI systems clearer contextual relationships to work with.

The problem gets worse as B2B websites grow

I see this particularly often with established B2B organizations.

Websites rarely begin as a mess. They become one gradually.

A new product launches, so a landing page gets created. Marketing builds a campaign microsite. An acquisition introduces another domain. A new team creates slightly different terminology for an existing capability. Five years of blogging produces hundreds of overlapping articles.

Eventually, the business might have enormous amounts of content but surprisingly little clarity.

I've encountered variations of this across many of the organizations I've worked with. Part of the SEO work I've done with companies such as Oteemo has involved connecting technical capabilities, solutions and thought leadership much more closely with the company's ICP and go-to-market strategy.

That experience has reinforced something for me: more content doesn't necessarily create more authority if the relationship between that content is unclear.

In some cases, consolidation may be more valuable than creation.

Your website should communicate relationships, not just information

This is where information architecture becomes particularly interesting for AI search.

Let's say a company wants to become more visible for enterprise AI security.

Creating a single optimized page called "Enterprise AI Security" might help, but I would want to understand the wider ecosystem around it.

Does the website explain the security problems associated with enterprise AI? Are there relevant products or services? Are there technical resources? Is there original expertise from subject-matter experts? Are there customer examples? Does the company discuss governance, implementation and adjacent use cases? Are these resources meaningfully connected?

That creates a much stronger information environment than simply publishing another SEO page.

I think about this as moving from pages to relationships.

Instead of:

Keyword → Page

We should increasingly be thinking:

Entity → Problem → Expertise → Solution → Evidence → Outcome

That doesn't mean every website needs an enormous taxonomy or hundreds of supporting pages. In fact, often the opposite is true.

The objective is clarity.

Internal linking becomes part of the story

Internal linking is another area where traditional SEO thinking and AI discoverability begin to overlap.

I've always considered internal links important because they help distribute authority, establish hierarchy and help users and crawlers discover related content.

But strategically, internal links also communicate relationships.

A thought-leadership article about AI governance should logically connect with the company's AI solution. A use-case page should connect to relevant customer evidence. An industry page should connect with the solutions applicable to that industry.

The link shouldn't exist because an SEO tool told us the page needs three more internal links.

It should exist because the relationship makes sense.

That creates a website that behaves more like a knowledge system than a collection of landing pages.

Schema helps, but it cannot fix an unclear website

Structured data naturally enters this conversation because schema can help machines understand entities and relationships.

I absolutely believe schema has a role in modern SEO and AI discoverability.

But schema should describe a clear information architecture, not compensate for the absence of one.

If the underlying content is duplicated, terminology is inconsistent, services overlap and the site's hierarchy doesn't make sense, adding structured data isn't going to magically create clarity.

I think the sequence should look more like this:

That ordering matters.

This is also why SEO should be involved before a website launches

Information architecture is difficult to retrofit.

If SEO and discoverability are only considered once the sitemap has been approved, the design is complete and development is almost finished, some of the most important decisions have already been made.

This is why I increasingly advocate for SEO to be involved in website strategy from the beginning.

Keyword research can tell us how people describe their problems. Search data can expose categories the business hasn't clearly represented. Sales conversations can reveal important use cases. Analytics can show where existing journeys break down.

Together, those insights can help shape the architecture itself.

The result isn't simply a website that is "SEO friendly."

It is a website that more accurately represents how the market understands the business.

AI discoverability starts with being understandable

I've written previously about AI Search Visibility and the need to think beyond individual rankings toward visibility across a wider set of search surfaces.

But before worrying about whether ChatGPT, Google AI experiences or another platform mentions your company, I think businesses need to answer a more basic question:

Have we made it easy for machines to understand what we actually do?

There won't be a single optimization that solves that.

It comes from consistent positioning, clear entities, logical site architecture, useful content, meaningful internal relationships, credible evidence and technical foundations that allow that information to be discovered.

In other words, some of the most important work we can do for AI search looks surprisingly similar to good SEO, good UX and good website strategy.

The difference is that we're no longer designing architecture purely to help somebody navigate a website.

We're designing an information environment that helps both people and machines understand the business.