On June 12, Google released OKF, an open specification for structuring business knowledge so AI agents can read it directly. Here is what it means for professional services firms.
What You'll Learn
- What the Open Knowledge Format is and where it came from
- Why professional services firms are a strong fit for it
- What an OKF knowledge bundle looks like
- How we are using it for clients and what the process involves
If you work in professional services, this post is worth five minutes of your time. Law firms, accounting practices, consulting firms, financial services companies. What I am about to describe came out of Google Cloud earlier this month, and most people in those industries have not seen it yet.
On June 12, Google released something called the Open Knowledge Format, or OKF. It is an open specification for structuring business knowledge so AI agents can read it directly. That description sounds technical, but the idea behind it is straightforward.
AI assistants are now answering questions for your clients and prospects before they ever visit your website. Someone asks ChatGPT which accounting firms in Boston handle multi-state tax compliance, or which law firms have handled a particular type of business litigation. The AI pulls together an answer from whatever it can find and understand about the firms in that space.
Most professional services websites are not built for that. They were built for human visitors and Google search. AI agents read differently, and most sites do not give them enough structure to work from accurately.
OKF is Google's answer to that problem.
What OKF Is
The format is intentionally simple. It is a collection of short markdown files, each one describing a single concept about your business. One file for each service you offer. One for each type of client you serve. One for each common question you field. One for key processes or workflows.
Each file has a small structured header identifying what the document is about, and then a plain-language body. The files link to each other, so an AI agent reading about one service can follow links to see which client types it applies to, or which questions come up most often around it.
That granularity is what makes the format useful to AI agents. They do not need to load your entire website. They find the specific concept they need and follow links from there.
What goes in a knowledge bundle
- One concept file per service or sub-service
- One per client type or industry served
- One per common client question
- Playbook files for key processes
- Cross-links between related concepts
- Citations back to your live website pages
What AI agents get from it
- A structured map of your business
- Clear descriptions of each service
- Which clients each service applies to
- How your services connect to each other
- A freshness signal from the update log
Why Professional Services Firms Are a Strong Fit
Professional services firms are, in my experience, the businesses most underrepresented in AI-generated answers relative to the quality of work they do. A few reasons for that.
Most professional services websites are heavy on credentials and light on explanation. They tell you the firm has 30 years of experience and a long list of practice areas, but they do not explain what those practice areas involve, who they are for, or how the firm approaches them differently from competitors.
AI agents need the explanation, not just the credential. They are trying to answer a specific question for a specific person. If your website does not clearly describe what you do and for whom, the AI has to guess or skip you.
OKF gives you a structured way to provide that explanation in a format AI agents are built to read. A law firm, for example, could have concept files for each practice area, sub-concepts for specific case types within each practice, files for the types of clients they typically represent, and files explaining how the intake and engagement process works.
The result is that an AI agent researching law firms for a specific type of matter can find the relevant concept, understand what the firm does in that area, see which client profiles fit, and link back to the relevant pages on the live website for more detail.
What We Are Building for Clients
We have been deploying OKF bundles for clients over the past week, starting with the most AI-visible work we are already doing. The process starts with a content audit: we identify the distinct concepts your business covers across your existing pages, then extract and structure them into concept files.
For a typical professional services firm, an initial bundle might include 25 to 40 concept files. Products or service lines get their own sub-concepts. Industries or client types get files. Key processes get playbook files that walk through how work gets done. Common client questions get FAQ-style concept files.
We then link the OKF bundle from the existing agents.md and llms.txt files already on these sites, so AI agents visiting the site have a clear path from the high-level introduction directly into the detailed knowledge graph.
Why the Timing Matters
OKF was released five days ago. The WordPress plugin for auto-generating OKF bundles is currently awaiting approval in the WordPress plugin directory. Essentially no small or mid-size businesses have implemented this yet.
That will change. Google has committed to supporting OKF across its AI and data products. The format is open and vendor-neutral, which means other platforms will adopt it. Early implementation means AI agents have an accurate, structured picture of your business before the standard becomes widespread and competitive.
For professional services firms specifically, where trust and expertise are the core differentiators, being accurately represented in AI-generated answers is not a small thing. Getting it right early is worth the effort.
If you want to understand how AI systems are currently reading your firm, the fastest test is to ask a few AI assistants directly. Ask ChatGPT to describe your firm, your practice areas, and what types of clients you typically serve. Gaps and errors in those answers are a reliable indicator of where structured signals are missing.
Ready to build an OKF knowledge bundle?
We start with an AI visibility review of your current site, then scope, build, and deploy the bundle within two to three weeks.
Common Questions About OKF
Is OKF the same as llms.txt or agents.md?
No, but they work together. llms.txt and agents.md act as a short introduction that points AI agents to your content. An OKF bundle is the structured content itself, the detailed knowledge an AI reads once it has been directed there. They are different layers of the same infrastructure.
Do you need technical resources to implement this?
Not for WordPress sites. Everything deploys through your existing site without server access or developer involvement.
Does OKF replace SEO?
No. Traditional search rankings are still the strongest predictor of AI citation. OKF sits on top of your existing SEO and extends your visibility into the AI layer rather than replacing anything.
How is OKF different from adding schema markup?
Schema markup tells search engines and AI systems what a specific page is about. OKF creates a navigable graph of your entire business knowledge that AI agents can traverse independently. Both matter, and they complement each other.
Conor McDonough — Managing Partner, Scribendi Digital Marketing
Conor is Managing Partner of Scribendi Digital. For more than 12 years he has been designing, building, and optimizing websites for search from top to bottom. He is a constant student of search and remains on the cutting edge of digital content marketing.
