Your Website Isn't Just For People Anymore

July 08, 2025

Search used to be about human intent (we’re still focused on ‘search intent’ afterall). Writers write blog posts and landing pages so they show up in results, convert traffic, and build relationships. But the next wave of internet visitors won’t be people, they’ll be bots.

LLM-powered agents like Agent S2 are starting to crawl, read, and interpret web content directly, bypassing traditional search interfaces altogether. Rather than focus on backlinks or keyword density, AI crawlers prioritize semantic embeddings—vectorized representations of meaning, context, and even geographic information.

Its a change that opens up a new front in SEO, where your content is judged by how well it informs models rather than how well it ranks for humans.

Here is what to expect in the near future

The Linkedin post that goes with this tweet is set in 2030, but the changes below might occur in half that time. Here is what to expect

1. The rise of AI agents as crawlers

Agent S2 is an example of the next-gen LLM bot designed to read and act on web content. They don’t just index sites, they interpret them, generating summaries and judgments based on page content. They can also perform actions based on them. Best of all, they learn as they go.

2. Embeddings Are the new index

Unlike traditional crawlers, AI agents generate embeddings—numerical representations of meaning. LLMs may derive geospatial embeddings as well, pulling signals from location references, local business data, or regional terminology. The implications are enormous and not fully grasped yet.

AI agent will power personalize answers (“best sushi near me”) even when sites aren’t geotagged in a structured way. But they’ll also pull a ton of other data that they may be able to exploit as well.

3. Introducing Semantic SEO

Semantic SEO means that is content optimized not only for keyword matching, but also for embedding quality and semantic richness.

Examples include:

• Structuring FAQs in natural language for better summarization.
• Using latent entity mentions (people, places, brands) to signal authority.
• Including clean metadata and full sentence structures to help LLMs generate accurate embeddings.

4. The implications for marketers

• Traffic might not be the goal anymore—being cited or embedded in AI outputs could become the new KPI.
• Brands may need bot-specific versions of their content (like JSON-LD for structured data, but semantically richer).
• Some sites might even start obfuscating or watermarking content to prevent unauthorized training.

What comes next

There will be a gradual phase-in of having AI agents rather than humans becoming your site’s most frequent visitors. But optimizing for them doesn’t mean abandoning human readers. It means understanding that meaning now matters as much as metadata, and your content’s structure, clarity, and context all feed into how it’s interpreted by models like Agent S2.

Semantic SEO isn’t not just tweaking traditional search strategy. It’s a signal of what’s coming. The sites that adapt early won’t just show up in search results. They’ll be woven into the answers and decision-making flows of future AI-driven interfaces (even if not holographic).