Website design and build
A site built to be found, from the start.
We build it from the start in a structure both AI and search can read. Available in Japanese, English and Korean.
What we do
What we do
A good looking site and a findable site are two different things. Can an AI crawler read the body in the first response? Are the company, the products and the people connected by structured data? Are the language versions tied together with hreflang?
AI crawlers read it in the first response.
We render the body on the server and put llms.txt, sitemap and robots in place, so an AI crawler reads the body in the first response.
Tie company, product and person into one.
With Organization, Person, Service and FAQ structured data, we tie the company, the products and the people into a single entity.
Connect the three languages.
We tie Japanese, English and Korean together as one entity with hreflang. This is not the same as translating pages.
Publish, and get cited.
We build the publishing machinery (the blog) as part of it, and design from the start the flow where what you write is read by search and AI, and cited.
Process
How it goes
The order is the same whichever service you take.
Research
We gather the searches and questions that look like yours, and see who is on top and what the AI is citing.
Trace the sources
We map out what gets cited and what gets rated, and identify the outlets lifting your competitors.
Prepare
Working back from the result you want, we decide what to fix and what to write.
Build, then measure
We ask the same question again and watch what moved. We keep going until you are in the answer.
FAQ
Frequently asked questions
We have a site, so why does ChatGPT not know the company?
Either the AI crawler cannot read the body, or the company, products and people are not tied into one entity. A site rendered only in the browser has no body in the crawler’s first response. Without structured data, an AI does not know how the company and the product relate.
What does it take to let AI crawlers (GPTBot, ClaudeBot) read the site?
Three things. Allow the AI crawlers in robots, render the body on the server so it is in the first response, and explain the company and products directly in llms.txt. This site itself implements all three.
The agency that built our site left out structured data. Can it be added later?
It can. We start with Organization, Person, Service and FAQ structured data and tie company, product and person together with sameAs. The machine readable layer can be added without touching the design.
Can Japanese, English and Korean sites be built as one site?
They can. Each language version points at the others with hreflang, and structured data makes them one entity. We built miruai.jp that way across three languages, and we build with the same design.
With a Korean or US head office site, should the Japanese site be on a separate domain?
It depends. Sitting under the head office domain works, and a separate domain for Japan works. Either way, the point is tying it to head office as one entity with sameAs in the structured data. We look at where you are and propose the shape that helps most in Japanese search and AI.
Why do you build with Next.js?
Because the body can be rendered on the server. Search engines and AI crawlers read the body in the first response, and the site is fast for people at the same time. This site and miruai.jp share the same setup.
How is this different from a design agency?
We build the structure that gets read, not the way it looks. We design the page people see and the structure search and AI read at the same time. Then we measure how the site we built is being found in search and AI with our own tools, and fix it.
Can you improve an existing site?
Yes. We look at where you are first and tell you what is missing in the structured data, the rendering or the language design.
What happens after the build?
We keep measuring your discoverability in search and AI, decide what to fix, and continue. The scope of that work is given to you in writing.
Contact