Google Launches Open Data Format, an AI Normal

Google has launched Open Data Format (OKF), a brand new open specification designed to standardize the way in which content material is packaged and shared with AI brokers. Google Cloud Tech’s X put up describes it as a “vendor-neutral” commonplace, constructed to be readable by each people and machines with out requiring new instruments or software program to implement.

What’s Open Data Format?

OKF v0.1 makes use of a listing of markdown information with YAML frontmatter to determine standardized conventions, making content material extra suitable between producers and AI brokers. In response to Google, a bundle of OKF paperwork is:

• “Simply markdown:” each file is obvious markdown that opens in any editor, renders on GitHub, and will be listed by commonplace search instruments
• “Simply information:” a bundle ships as a tarball, lives in any git repo, and mounts on any filesystem
• “Simply YAML frontmatter:” a brief listing of queryable fields covers sort, title, description, useful resource, tags, and timestamp

Google has emphasised the simplicity of the format. There isn’t any compression scheme, new runtime, or required SDK concerned. 

For now, the reference implementations and tooling heart on Google Cloud, however the spec itself is open and vendor-neutral, printed on GitHub underneath the Apache 2.0 license. OKF v0.1 will evolve as producers and customers study what information representations brokers really want.

Why this issues for entrepreneurs

It is nonetheless early. OKF is an inside information format for AI brokers, not a search rating or web-publishing sign, and it is not tied to Google Search, YouTube, Maps, or different client merchandise.

Google’s argument is that OKF can pull solutions from in any other case incompatible sources. Its personal examples are inside and technical: desk definitions, metrics, and runbooks. For advertising and marketing groups, the parallels are inside too, like model information bases, reporting, gross sales enablement, and coaching.

OKF is not the one commonplace pointed at AI brokers, and the variations are straightforward to confuse. An XML sitemap lists the pages in your web site. An llms.txt file factors crawlers towards your most helpful content material. OKF goes a step additional and palms over the information itself, packaged as information an agent reads immediately. The distinction that issues is path: llms.txt and sitemaps face outward, towards the crawlers visiting your web site, whereas OKF faces inward, towards a corporation’s personal brokers.

For many entrepreneurs, none of this requires motion at present. A broadcast OKF bundle will not transfer your rankings this week or subsequent. Guarantee your content material is clear and well-structured, watch how the usual develops, and ensure that AI crawlers can already attain the content material you publish. If agent-facing requirements catch on the way in which sitemaps and schema did, that foundational work pays off.

The right way to put together with Semrush

Website Audit flags whether or not your web site is obstructing AI crawlers and surfaces technical points that may hold your content material out of AI-generated solutions, together with a lacking llms.txt file. It runs on each Semrush plan, so it is a low-lift place to start out.

To see how that work interprets into visibility, the AI Visibility Toolkit tracks how typically your model will get cited throughout ChatGPT, Perplexity, Google AI Mode, and Gemini.

For enterprise groups, Semrush Enterprise AIO goes deeper, with AI visibility monitoring throughout extra LLMs and Crawler Profiles that simulate how bots like ChatGPT and Googlebot entry and interpret your web site.