TL;DR: Google just published the Open Knowledge Format (OKF), an open standard for organizing business knowledge into markdown files that AI agents can read without translation. While OKF was built for internal data teams, the underlying idea applies directly to your website. If you structure your site's core information (services, locations, FAQs, expertise, processes) the way OKF recommends, paired with proper schema markup, you give AI search engines a cleaner, more trustworthy version of your business to cite. This matters because AI Overviews, ChatGPT, and Perplexity are increasingly answering customer questions before anyone reaches your homepage. The businesses that organize their knowledge clearly are the ones getting cited.
On June 12, 2026, Google quietly published something that most business owners will never hear about, but that has real implications for how your business gets found.
It's called the Open Knowledge Format, or OKF. On the surface, it sounds like a developer tool. It is a developer tool. But underneath the technical language is a simple idea that every business owner and marketing director should understand: AI systems work better when information is organized in a predictable, structured way they can read without guessing.
Businesses should Care!
Businesses should care about OKF's underlying logic because it describes exactly what's missing from most websites right now: a clear, typed, linked structure that an AI system can read without guessing. Right now, when ChatGPT or Google's AI Overviews try to answer a question about your business, they're piecing together fragments from your homepage copy, a few service pages, maybe an outdated directory listing, and whatever else they can find, much the same way Google described internal teams piecing together scattered, mutually incompatible surfaces to answer a simple question. The fix isn't complicated. It's making sure every page on your site clearly states what it is, how it connects to the rest of your business, and what question it answers, using schema markup as the machine-readable layer underneath your normal copy. That's the same principle OKF formalizes for internal data: structure removes ambiguity, and removing ambiguity is what gets you cited instead of skipped.
That idea isn't new to SEO & SEO is NOT DEAD!
But Google just formalized it, and the formalization tells you something important about where search is heading. If you've been wondering why your competitor with the worse website keeps showing up in AI Overviews and you don't, this is part of the answer. Not the OKF format itself necessarily, but the principle behind it: structure beats hope.
What Google Actually Said....
Google's framing of the underlying problem is worth sitting with for a second. They describe how, inside most organizations, information that foundation models use lives in highly fragmented systems including metadata catalogs, wikis, code comments, and the heads of a few senior engineers. When an AI system needs an answer, it has to assemble that answer from scattered, mutually incompatible surfaces.
Now replace "organization" with "your business" and "AI agent" with "ChatGPT answering a customer's question about your services." Same problem. Different scale.
Why This Matters for Your Website, Not Just Data Teams
OKF was built for internal company data: database schemas, metrics definitions, runbooks. But the design principles behind it are exactly the principles that determine whether AI search engines understand your business correctly.
Google describes the core philosophy this way: a good knowledge format should be something anyone can produce without an SDK, anyone can consume without an integration, that survives moving between systems, and that's readable by humans and parseable by agents with no translation layer.
That description applies almost word for word to what your website should be doing right now. Your homepage, your service pages, your FAQ section, your about page. All of it is "knowledge" about your business that AI systems are trying to parse, summarize, and cite. The question is whether you've made that easy or whether you've left it to chance.
Here's the part that should get your attention if you run a business that depends on showing up when people search for what you do. Research on AI search behavior found that 65% of pages cited by ChatGPT include structured data, and separate analysis found that content with proper schema markup has a 2.5x higher chance of appearing in AI generated answers, with sites that have complete structured data seeing up to 40% more AI Overview appearances.
A 2.5x difference isn't a rounding error. That's the gap between being part of the conversation AI is having with your customers and being invisible in it.
The Three Things OKF Gets Right (And Your Site Probably Doesn't)
You don't need to implement OKF on your website. It's a backend data format for AI agents querying internal company systems. But the three principles behind it map directly onto how you should be thinking about your site's content architecture.
Every piece of content needs a clear "type"
OKF requires exactly one thing of every concept document: a type field. Everything else is flexible, but the system always knows what kind of thing it's looking at.
Most business websites don't do this. A page about your services, a blog post, an FAQ, and your about page all read as undifferentiated text to a crawler unless you tell it otherwise. Schema markup is how you tell it otherwise. Marking your service pages as Service, your FAQ as FAQPage, your business details as LocalBusiness or Organization gives AI systems the "type field" that OKF considers non-negotiable.
Research from early 2026 found that in a controlled experiment with three nearly identical pages, only the page with well-implemented JSON-LD schema appeared in a Google AI Overview, and it also achieved the highest organic ranking. The page with no schema never even got indexed.
Read that again. Not "ranked lower." Never got indexed.
Knowledge should link to other knowledge
OKF documents link to each other with normal markdown links, turning the directory into a graph of relationships that is richer than simple file system hierarchy.
This is the part of "knowledge graph" that most websites get backwards. A knowledge graph isn't a fancy diagram. It's the web of relationships between entities: this business offers this service, in this location, for this type of customer, backed by this credential, addressing this problem.
If your service pages don't clearly connect to your location pages, your about page doesn't connect to your team's credentials, and your FAQ answers don't connect back to the services they're describing, you don't have a knowledge graph. You have a pile of disconnected pages that happen to share a navigation bar.
Entity-level schema (Organization, Person, LocalBusiness with sameAs links to verified profiles) is how you build those connections in a way machines can follow. Industry analysis of post-update AI search behavior notes that Organization and Person schema with sameAs identifiers enable AI to resolve the publishing entity against Knowledge Graph records, and resolved entities receive higher trust scores in AI answer generation.
The format should survive without a translation layer
OKF's whole premise is that knowledge should be readable by humans and parseable by agents, the same file, no translation layer.
For your website, this means the information AI systems need to answer questions about your business shouldn't be buried in marketing copy that requires interpretation. If a customer asks an AI assistant "does this company serve my area" or "what does this service actually include," the answer should exist somewhere on your site in a form that's both readable to a human and explicitly tagged for a machine.
This is the difference between writing "we proudly serve communities across the region" (sounds nice, tells a machine nothing) and having a LocalBusiness schema entry with your actual service area, plus a page that names the towns and cities by name. One requires interpretation. The other doesn't.
What's Actually Changing in How AI Reads Your Site
It's worth understanding the shift Google is signaling here, because it's bigger than one format announcement.
Analysis following Google's March 2026 update found that the update didn't diminish the value of structured data, it changed what structured data is valuable for, shifting schema from a SERP display trigger to an AI trust and entity verification signal. The same analysis noted that Google's internal documentation confirms AI Mode source selection considers structured data quality alongside PageRank signals, content freshness, and query relevance.
In other words: structured data used to be about getting a star rating to show up in search results. Now it's about whether an AI system trusts your business enough to mention it by name when someone asks a question.
Google's own confirmation, reported from their I/O 2026 developer sessions, states it plainly: schema.org structured data remains valuable for rich results even though no special markup is required for AI responses specifically, and schema is read as a trust signal, not a display trigger.
OKF is Google telling its own engineering teams: stop making AI guess. Give it clean, typed, linked, well-organized information, and it will use it correctly. That's the same message your website needs to receive.
What This Looks Like in Practice
You don't need to rebuild your site or learn a new file format. What you need is an honest audit of whether your site's information is structured the way OKF assumes good knowledge should be structured.
Start with your core entity. Does your homepage have Organization schema with accurate name, address, sameAs links to your verified social and directory profiles, and a clear description of what you do? This is your knowledge graph's anchor point. Everything else connects back to it.
Then look at your services. Each one should be its own page, marked up as a Service or Product where appropriate, with a clear type, a clear description, and links to the location pages and FAQ entries that relate to it.
Then your FAQ. This is the section most businesses get wrong, usually by skipping it entirely or burying generic questions that don't reflect what people actually ask. FAQPage schema, paired with questions phrased the way real customers search (not the way your team talks internally), gives AI systems direct, citable answers.
Finally, look at the connections. Click through your site the way an AI crawler would. Does your about page connect to your service pages? Do your service pages connect to your location and team pages? Or does everything just point back to "contact us"?
This is the kind of structural work that doesn't show up in a flashy before-and-after screenshot, but it's the foundation everything else sits on. It's also exactly the kind of work that gets skipped when an agency is focused on running ads instead of fixing the underlying visibility problem.
Frequently Asked Questions
Do I need to actually use Google's Open Knowledge Format on my website?
OKF is designed for internal company data systems and AI agent infrastructure, but you can adopt it to your website. The relevant takeaway for your site is the underlying principle: organize your business information with clear types, clear relationships, and no ambiguity, using schema markup as the implementation tool.
What's the difference between a knowledge graph and just having a website?
A website is a collection of pages. A knowledge graph is the structured web of relationships between your business, its services, locations, people, and the questions customers ask, expressed in a way machines can follow without guessing. Most websites have the first without the second.
Will adding schema markup guarantee my business shows up in AI Overviews or ChatGPT answers?
No, and any agency that promises that isn't being straight with you. Structured data increases your eligibility and helps AI systems trust and accurately interpret your content, but it doesn't guarantee a citation. It removes ambiguity. The rest still depends on content quality, relevance, and authority.
Which schema types matter most for a small or local business?
Organization or LocalBusiness for your core entity, Service for what you offer, and FAQPage for common customer questions are the foundation. Article or BlogPosting schema matters too if you publish educational content regularly.
How long does it take to see results from improving site structure this way?
Structural changes like schema markup and internal linking typically show measurable movement within 30 to 60 days, based on observed patterns following structured data implementation. It's not instant, but it's also not speculative. It's foundational work that compounds.
The Real Takeaway
Google didn't publish OKF for small business owners. They published it for engineering teams building AI agents inside large organizations. But the announcement is a window into how Google's own systems think about knowledge, and that thinking flows downstream into how AI search treats your website too.
The businesses that will be cited, recommended, and surfaced by AI search aren't necessarily the ones with the most content. They're the ones whose content is organized so clearly that a machine doesn't have to guess what anything means. If you want a second set of eyes on whether your site's structure is helping or hurting your AI search visibility, book a growth audit call with Rakos Media. No pitch, just diagnostics.
Sources
- Introducing the Open Knowledge Format — Google Cloud's announcement and technical description of OKF, including its design principles and the problems it addresses.
- Schema Markup for AI Search: 65% of AI-Cited Pages Use It — Data point on structured data prevalence among AI-cited pages.
- Schema Markup Statistics 2026 — Statistics on the citation and visibility impact of complete structured data implementations.
- Schema Markup After March 2026: Structured Data Update — Analysis of how Google's March 2026 update shifted the role of structured data toward AI trust and entity verification.
- Structured Data After I/O 2026: Schema Cheat Sheet — Google's I/O 2026 confirmation on how schema functions as a trust signal for AI Mode.
- Structured Data for SEO: A Guide to Schema Markup in 2026 — Controlled experiment showing the indexing and AI Overview impact of JSON-LD schema versus no schema.
