You've spent 20 years optimizing for humans: Now it's time to optimize for AI

For the past 20 years, the travel industry has built websites for both humans and search engines. We've created multi-destination sites with schema markup, JavaScript frameworks and structured URLs. We've watched every Google algorithm update and adjusted accordingly.

But now there is a major change in how content is found online. Large language models (LLMs) are changing how the web is indexed, processed and served to users—and most travel companies haven't caught on yet. 

While we have become good at search engine optimization (SEO) for Google and humans, artificial intelligence (AI) reads and processes web content differently. Travel websites, especially those with dynamic, multi-destination content, need to optimize not just for traditional search engines but for AI systems that will increasingly direct travelers to our content.

AI doesn’t read your website like Google does 

Traditional search engines like Google have specific ways of crawling and indexing your content:

• They render JavaScript (eventually)
• They understand schema markup and structured metadata
• They consider backlinks, domain authority and page speed
• They associate your content with your domain

LLM-based indexers that feed content to AI systems like ChatGPT, Gemini or Claude work differently:

• Many don't render JavaScript, they read raw HTML only
• Most ingest unstructured text; schema and structured data often get ignored
• Content becomes part of a general training set with no guaranteed attribution
• They don't necessarily tie content to source URLs in the way Google does

This means the SEO practices you have used for two decades might be nearly invisible to AI systems that will increasingly direct travelers to our content.

Five tactical ways to optimize for AI

1. Serve a crawler-friendly version of your site

AI crawlers are not as advanced as Google's rendering systems. Creating simplified versions of your content can improve how AI systems process your pages:

• Use pre-rendering tools like Rendertron or Prerender.io
• Consider serving a simplified version of your site to known AI crawler user agents with a clear HTML structure, easy-to-follow links and page hierarchy and no JavaScript dependencies

2. Rethink your content strategy for AI readability

People interact with AI differently than they do with search engines:

• Focus on conversational Q&A: SEMrush reports that LLM queries average 23 words, versus two to four words in traditional search. Create specific FAQ sections that match how people ask AI systems questions. 
• Make your brand unmissable: Unlike Google, which ties content to domains, AI systems pull from a text corpus where your brand name might be the only identifier. Create content with plenty of branded sentences, like "Travelandco offers guided tours through 12 regions of Italy," and "Travelandco's hiking excursions include equipment and local guides."

Consider how these might introduce your brand to new customers through AI responses. When someone asks Gemini, "What tour companies offer guided experiences in multiple regions of Italy?" or asks Claude, "Which hiking companies provide equipment with their tours?", your brand could appear in responses to users who had never heard of you before.

• Simplify your site structure: Flatten and streamline your internal linking so AI crawlers can easily discover all of your content.

3. Understand the new value signals

What matters for AI visibility is different from traditional SEO:

• Text mentions may outweigh backlinks: Since many LLMs ingest text rather than analyzing link structures, simple mentions of your brand name might carry more weight than carefully built backlinks. This is because typical backlinks do not usually include branded mentions in their anchor text unless they are from press releases or news coverage.
• Schema becomes just text: Unless there's a retrieval-augmented generation (RAG) system specifically built to interpret your schema markup, it's just more text in the training corpus—assuming the AI systems even read your schema.

4. New measurement challenges

Tracking visibility in AI systems isn't straightforward, but tools are emerging:

• SEO platforms like Semrush and Ahrefs are beginning to offer insights into AI visibility
• Startups focused specifically on AI optimization are launching measurement tools
• Google Search Console will likely add AI visibility metrics eventually

Until then, consider testing. Ask AI systems specific questions about your offerings, and see if your brand appears in responses.

5. Looking ahead: AI crawling will evolve rapidly

What works today won't necessarily work tomorrow:

• AI crawlers will get better at JavaScript rendering
• RAG systems will eventually give structured data renewed importance
• Knowledge graphs will help AI systems understand entities and relationships
• Attribution systems will improve

If I were building for AI from scratch

If I were building a travel website today with AI visibility as a primary goal, I'd create a parallel version of the site specifically for AI crawlers. Yes, this is basically like the early days of black hat SEO all over again, but hey, nobody is saying there is anything wrong with this approach yet. This would include:

• Plain HTML with minimal styling
• Clear information architecture
• Comprehensive FAQs that match conversational queries
• Content with frequent brand mentions and clear value propositions
• Simple internal linking system that makes all content discoverable

In more technical configurations, this can be handled through user agent routing that serves different content based on the crawler that is visiting your site.

This approach wouldn't replace your existing site, it would add to it, making sure you're visible across all discovery channels.

The stakes for travel brands