Machine Legibility
How do AI tools read and describe us? AI search and agents now sit between products and buyers, and they describe you from what they can read. Four layers keep that description accurate: a consistent description, crawlable content, documentation machines can read, and genuine mentions elsewhere.
Four layers between you and the machine.
- Layer 1One descriptionWho you are, worded the same way everywhere.
- Layer 2Crawlable contentPages a crawler can read without running scripts.
- Layer 3Machine-readable documentationStructured data, a sitemap, and docs written to be parsed as well as read.
- Layer 4Genuine mentionsOther sites describing you in their own words, because they chose to.
This is not a promise of AI citations. Nobody controls what a model says. What you control is whether what it reads about you is accurate and consistent.
What each layer needs
One description
A single sentence for who you are, used on your site, your profiles and your press kit.
Ask the main AI assistants who you are, and compare their answer with your sentence.
Crawlable content
The words live in the page itself, not behind a script or inside an image.
View the page with scripts turned off. Whatever disappears may be invisible to crawlers that don't run them.
Machine-readable documentation
Structured data that says what each page is, a sitemap of every public page, and product docs written so a machine can pull out the facts.
The structured data validates, and the sitemap matches the site.
Genuine mentions
Other people describing you, in places you don't control.
Mentions you earned. Bought or manufactured ones do more harm than good.
This site, layer by layer.
Own work
One description
Every page describes us the same way: an AI-native product and venture studio based in Ahmedabad, India. We pair the name with Ahmedabad on purpose, because another company has a similar name.
Crawlable content
Every page is plain HTML. The framework diagrams, including the one on this page, are text rather than images, so they read the same to people, screen readers and crawlers.
Machine-readable documentation
Structured data describes the organisation and where each page sits in the site, a sitemap lists every public page, and our robots file lets search engines, AI search and AI assistants read all of it.
Genuine mentions
The one layer we can't write ourselves. We publish work worth citing, like this library, and leave the rest to other people.
Where it goes wrong
- Misread
Expecting citations
Why This makes the description accurate. It can't guarantee a mention.
- Misread
Writing for machines instead of people
Why The same page serves both. Copy written only for crawlers reads badly to the buyers it's meant for.
- Misread
A different description for every channel
Why Five slightly different sentences leave the machine to average them.
In the Journal: Teardown: our old website
Published 24 September 2026. Graylemon original, from our third conviction: machine is the new audience.