When Was the Last Time a Human Saw Your Site?

Cloudflare has announced the launch of “Kitesurf,” an AI agent-specific browser. While it may seem unremarkable at first glance, this news has the potential to upend the very foundations of the web.

What does this mean? Until now, websites were designed to be “opened and read by humans”. SEO strategies, design, and user flow have all been built with the “human eye” in mind. However, Kitesurf is a browser specifically for AI agents to navigate the web. Humans are not involved. AI will autonomously open sites, read information, make judgments, and take the next actions.

In other words, it may no longer be humans who are “reading” your company’s homepage.

What is Kitesurf? — A Shift in Cost Structure

The key point of Kitesurf is that it consumes significantly fewer computational resources compared to traditional headless browsers (which process web pages without displaying them). According to Cloudflare’s announcement, it can greatly reduce the resources needed compared to conventional browser automation.

This implies that the cost of AI agents crawling the web will dramatically decrease.

What happens when costs go down? The volume explodes. AI-driven web crawling, which has been limited by technical and cost constraints, will suddenly become commonplace. Just as Google’s crawlers navigate the web, countless AI agents will soon be coming to read every site.

At the same time, the standardization of communication protocols between AI agents, such as Anthropic’s Model Context Protocol (MCP) and Google’s Agent2Agent Protocol (A2A), is rapidly advancing. Agents will collaborate, exchange information, and complete tasks. The need for humans to open browsers will undoubtedly decrease.

What Will Happen to Small and Medium-Sized Enterprises’ Websites?

Now, let’s get to the main point. What does this change mean for local small and medium-sized enterprises?

1. The Foundations of SEO Will Crumble

Until now, small and medium-sized enterprises have relied on SEO strategies to attract customers online. They would pay SEO firms between 50,000 to 300,000 yen per month, adjusting keywords to aim for higher rankings in Google search results. This amounts to an annual cost of 600,000 to 3,600,000 yen.

But what if users start asking AI assistants, rather than Google searches, “Find reputable construction companies in this area”? AI agents will evaluate and generate answers based not on Google search rankings but on structured data, review information, clarity of pricing, and frequency of updates.

Even if you rank first in SEO, it won’t matter if you’re not included in the AI’s responses. Conversely, even if you’re outside the top ranks in SEO, if the AI determines that “this information is reliable,” you may still be chosen. The very concept of ranking will dissolve.

2. An Era of ‘Structure’ Over ‘Appearance’

The typical cost for creating a website for small and medium-sized enterprises ranges from 500,000 to 2,000,000 yen, with a significant portion allocated to design. Attractive header images, animations, and stylish fonts are all aimed at capturing human attention.

For AI agents, none of this matters. What AI looks at are the HTML structure, structured data like Schema.org, meta information, and the logical composition of text.

To put it bluntly, a beautiful site that cost 1,000,000 yen for design may be less “readable” to AI agents than a 50,000 yen site built with a free WordPress theme that has properly configured structured data.

This actually presents an opportunity for small and medium-sized enterprises. A brand site that cost a large corporation tens of millions of yen and an information-focused site created by a small business for a few thousand yen can stand on the same playing field in the eyes of AI. The disparity in design will be nullified.

3. The Rules of Competition for E-commerce Sites Will Change

For local small and medium-sized enterprises operating e-commerce sites, the impact is even more direct.

If you instruct an AI agent to “find the cheapest product with these specifications,” the agent will crawl multiple e-commerce sites, comparing prices, specifications, stock, and shipping costs to return the optimal choice. Here, the crucial factor is whether the AI agent can accurately read the product information.

If product names are embedded in image text and not listed in the HTML, or if prices are dynamically generated without structured data, such sites will be treated as if they “do not exist” by AI agents.

Conversely, sites that structure product data using JSON-LD, provide real-time stock information, and expose API endpoints will be more likely to be selected by AI agents.

What is ‘AEO’? — The Steps to Take Are Surprisingly Simple

The term AEO (AI Engine Optimization) has started to emerge in contrast to SEO (Search Engine Optimization). It means optimizing for AI, but there’s no need to be swayed by buzzwords.

Let’s break down the concrete steps to take.

① Implement Structured Data (Estimated Cost: 50,000 to 200,000 yen)
Add structured data compliant with Schema.org to your site. This includes industry, location, business hours, product information, FAQs, and reviews. If you’re using WordPress, plugins can handle this easily. It’s not technically difficult.

② Set Up llms.txt (Estimated Cost: Almost Zero)
A specification called “llms.txt,” which is like an AI version of robots.txt, has been proposed. It’s a file that explicitly states which information you want AI to read. Just placing one text file will suffice. The cost is nearly zero.

③ Logical Structuring of Content (Estimated Cost: 100,000 to 500,000 yen)
Reorganize existing content into a format that is easy for AI to understand. Eliminate ambiguous expressions and describe facts and figures clearly. Increase the number of FAQs. This can be outsourced or done in-house using AI tools to organize.

④ Publish API/Data Feeds (Estimated Cost: 200,000 to 1,000,000 yen)
For e-commerce sites, enable the provision of product data via an API. This allows AI agents to directly access the data. There may be some technical hurdles, but many platforms like Shopify can handle this as a standard feature.

Overall, the minimum effort can start at 100,000 to 300,000 yen, and even a comprehensive approach can be initiated for 500,000 to 1,500,000 yen. For companies paying 3,000,000 yen annually for SEO outsourcing, simply reallocating a portion of that budget would be sufficient.

‘It’s Too Early’ Is Already Too Late

Many business owners might think, “AI agents don’t concern our industry yet.” However, the fact that Cloudflare has begun providing this infrastructure at an infrastructural level means that the speed of adoption is shifting from “years away” to “months away.”

Cloudflare is an infrastructure company that processes about 20% of the world’s web traffic. They are now offering the “AI agent-specific browser” as a standard feature. This represents a level of infrastructural change comparable to when Google created its search engine.

Remember, in the early 2000s when SEO started to be recognized as important, small and medium-sized enterprises that acted quickly outpaced large corporations to secure top search rankings and achieved significant results. The same is about to happen now with AEO.

So, What Should You Do?

Here are three things you can start doing tomorrow.

1. Test Your Own Site with AI.
Provide the URL of your site to ChatGPT or Claude and ask, “Tell me about this company.” If the AI cannot answer accurately, your site is considered “unreadable” by AI.

2. Check the Status of Your Structured Data.
Use Google’s “Rich Results Test” tool to check your site. It’s free and takes just three minutes.

3. Review Your Contract with SEO Firms.
Consider how much of your monthly SEO budget can be reallocated to “AI readiness.” If your SEO firm is not mentioning AEO at all, it may be time to consider changing providers.

The readers of your website are shifting from humans to AI. This is not a threat; for small and medium-sized enterprises, it’s an opportunity to bridge the gap with large corporations that they couldn’t compete with in design or advertising costs, by focusing on the quality and structure of information.

What’s important is to first view your own site through the eyes of AI. Just doing that will reveal what you need to do next.