Learn · Publishing

How do I publish company information for AI?

Short answer. Keep your company facts in one structured, approved source; publish them as clear, server-rendered pages with consistent terminology and structured data; add machine-readable feeds or an API where useful; and keep everything current and dated.

Last reviewed 24 September 2026

Why this matters

People increasingly ask AI assistants about businesses: what a firm does, where it is based, whether it handles a particular kind of work. Those assistants draw on what they can find and understand. If your information is scattered, inconsistent or hidden inside images and scripts, AI systems are more likely to describe you vaguely or wrongly.

You cannot control what any AI says. You can make accurate information easy to find and hard to misread.

Step 1: decide what is public

List the facts you want the world to know: services, sectors, locations, contact routes, team, opening hours, policies, frequently asked questions. Just as importantly, decide what is not public. Having an explicit list prevents internal details leaking into public pages.

Step 2: keep one source of truth

The most common problem is inconsistency: one address on the website, another on a directory, a third in an old PDF. Keep company facts in one structured, maintained place and publish from there, so every channel says the same thing.

Step 3: publish clear pages

  • Server-rendered HTML. Content that only appears after scripts run is harder for some crawlers and AI systems to read.
  • One page per topic. A page for each service, each location and each key question.
  • Plain definitions. Say what you do in a sentence near the top of the page.
  • Consistent terms. Use the same name for the same thing everywhere.
  • Semantic structure. Real headings, lists and tables rather than styled text.

Step 4: add structured data

Schema.org structured data, commonly embedded as JSON-LD, describes your organisation, services, locations and articles in a form software can parse. Search engines publish guidance on which types they support. It supplements good pages; it does not replace them.

Step 5: offer machine-readable routes

Depending on your audience, consider:

  • A sitemap so crawlers find every page.
  • A public feed or API for company facts, services or articles.
  • A public MCP endpoint that lets AI agents query approved information directly.
  • A short machine-readable summary file where emerging conventions suit your site. Treat these as optional; practices are still settling.

Step 6: keep it current

Date your content. Review it on a schedule. Stale information that AI repeats confidently is worse than no information.

Checklist

  • A written list of public facts, and of non-public ones.
  • One source for company information.
  • Server-rendered, well-structured pages.
  • Structured data for organisation, services and locations.
  • A sitemap and, where useful, an API or feed.
  • Review dates on every page.

How Kroy approaches it

In Kroy, company facts, services, locations, articles and FAQs are structured entities. You define a public projection of them, and Kroy Publish serves it through Channels: to your website through the API, to AI agents through MCP, and to other applications that need it. Update a fact once and every channel reflects it. Kroy provides the data and approvals; your website stays your own.