The Intelligent Web: Designing for Humans, AI Agents, and the New Buyer Journey

The website is entering its biggest transformation in three decades.

According to Darcy Kurtz, CMO at WP Engine, nearly every marketing team will need to rebuild their website in the next 24 months. Not because websites are broken, but because the environment around them has fundamentally changed.

“Websites are going through massive change,” Kurtz explains. “We see how traffic is changing. We see how bots are now your number one visitor and not humans.”

For decades, marketing teams built websites for one primary audience: people. Visitors searched, clicked, learned, compared, and converted. The website functioned as the education engine of the business.

Today, that model no longer tells the whole story.

Your Website Has Three Users Now

Kurtz outlines a critical shift. Modern websites must serve three distinct audiences:

1. The traditional human visitor
2. The LLM-educated buyer
3. AI agents that are not human at all

The traditional journey still exists. Many buyers still arrive through search and need education. That functionality cannot disappear.

But a rapidly growing segment of buyers now begins inside ChatGPT, Gemini, Perplexity, or Claude. By the time they reach a brand’s website, they are already informed.

“Buyers that start there have a very different experience and context by the time they get to your website,” Kurtz says. “They’ve probably already been told to buy you.”

Instead of researching, they are validating. Instead of comparing options across tabs, they are confirming that the recommendation aligns with reality.

Data at WP Engine reflects this shift. Visitors arriving from LLMs are converting at significantly higher rates than traditional traffic.

“We’ve now seen it in our own data that people that come in from an LLM are converting at six to seven times higher rate than people that come in through traditional ways,” Kurtz shares.

Then there is the third user: AI agents. These systems crawl websites to extract structured information and feed it back into LLM responses. They do not care about brand storytelling. They care about clarity, structure, and accuracy.

“All they care about is very structured data that is accurate and easy to gather and take back to their constituents,” she explains.

This creates a new reality. Websites must serve both emotional validation for humans and machine-readable structure for AI systems.

The Website Is Becoming a Brand Validation Layer

If education increasingly happens inside LLMs, what is the website for?

Kurtz frames it clearly. The website is evolving into a brand validation layer.

When a buyer arrives after receiving an AI recommendation, they are asking:

• Is this company credible?
• Does this brand feel trustworthy?
• Does the experience match what the LLM described?

“They want to understand this. Are you real? Are you credible? Is the LLM representing your brand in the right way?” she says.

That makes brand expression more important than many performance marketers have treated it in recent years. If a company positions itself as premium, the site must feel premium. If it claims enterprise authority, the experience must reinforce trust immediately.

At the same time, fast paths to conversion matter more than ever. In one example, WP Engine centered a “Buy Now” button in the hero section and saw double-digit growth in conversion rates.

“For the people that were coming in from LLMs to have that really quick fast pass, place dead center, not off center, shockingly made a huge difference,” Kurtz notes.

The lesson is not to remove educational journeys, but to layer in frictionless paths for buyers who are already ready.

From Keywords to Entity Mapping

One of the most significant conceptual shifts Kurtz describes is moving beyond traditional keyword optimization toward entity mapping.

Traditional SEO asked: what keywords should we rank for?

Entity mapping asks: what contexts, use cases, and attributes define our category and product in the real world?

Kurtz offers an analogy. Instead of describing a handbag only by color, size, and material, imagine how a buyer actually prompts an LLM:

“I am going on a yacht in the Mediterranean, and I need something that is going to go day to night.”

LLMs process contextual intent, not just product specifications. To appear in those results, websites must reflect those broader dimensions.

“You’re going to want all those pieces mapped out that make up what people might be asking for,” she says. “Do I have all that content on my site? And how do I get those elements so that when the LLM comes, they see it?”

Entity mapping typically includes:

• Core product attributes
• Target audience segments
• Use cases and environments
• Regional variations
• Emotional drivers
• Adjacent category terms

Once mapped, gaps become visible. Content strategy shifts from isolated blog posts to systemic coverage of a category’s full context.

GEO and the Power of Third-Party Influence

Another shift many marketers are underestimating is the role of third-party content in LLM responses.

“What we’re seeing right this minute is that the LLMs are prioritizing third-party content over your website content,” Kurtz explains.

That means generative engine optimization, or GEO, extends beyond on-site content. It includes:

• Identifying which publishers are cited in LLM responses
• Testing key prompts across major models
• Building partnerships with cited publishers
• Publishing structured, comparison-driven content externally

“You’re going to go after sources you never went after before,” she says. “The citations are not your top five publishers that you probably go after.”

This is not traditional PR. It is strategic influence over how AI systems synthesize information about your category.

The Back End Matters as Much as the Front End

While brand expression and content structure matter on the presentation layer, infrastructure is equally critical.

AI systems increasingly rely on protocols that allow them to retrieve structured data more efficiently. Marketing teams must work closely with web engineering to ensure sites are technically prepared for this machine-driven interaction.

“This is a front end and back end phenomenon,” Kurtz emphasizes. “You certainly on your website, how do you structure your data better or differently? But now on your back end, how do you make sure it’s ready for the agentic world?”

Rebuilding a website in this era is not just a design refresh. It is a structural evolution.

Rethinking Measurement in an AI World

Perhaps the most uncomfortable implication of all this change is measurement.

“In the last 30 years, we got hooked on this drug called attribution,” Kurtz says candidly. “And not only did we get hooked, our CFOs and our boards got hooked.”

If discovery increasingly happens inside LLMs, traditional click-based attribution models will weaken. Traffic may decline while revenue grows. The connection between source and sale becomes less linear.

Kurtz believes marketing leaders will need to return to more holistic measurement frameworks, combining:

• Revenue performance
• Brand awareness and perception data
• LLM citation tracking
• Sentiment analysis
• Impression share and visibility

“We’re going to have to go back to some of the classic ways of measuring performance,” she says.

It is less tidy than multi-touch attribution dashboards. But it reflects how buying decisions are now forming.

The Real Risk Is Standing Still

The website is not disappearing. It is evolving.

Education is moving upstream into AI systems. Validation, trust, and conversion are moving downstream onto brand-owned properties. Machines are reading content alongside humans.

Marketing teams that proactively adapt their content model, infrastructure, and measurement frameworks will shape how they appear in this new discovery layer. Those who do not may slowly lose visibility where buying decisions begin.

As Kurtz puts it, this is not about fixing something broken. It is about recognizing that “we now live in this hybrid world where there are multiple types of users that are coming to our site.”

For CMOs, the mandate is clear. Audit the experience. Map the entities. Structure the data. Strengthen the brand layer. Prepare the back end.

The rebuild cycle is coming. The only question is whether it happens strategically or reactively.