We are going to be straight with you. We are a new, independent agency, and we do not yet have a roster of named client case studies to show you. Publishing invented results, fake logos, or borrowed metrics would be dishonest, and it would also be exactly the opposite of what we sell: trustworthy, verifiable knowledge.
So instead of fabricating proof, this page does two useful things. First, it shows you the exact framework a real Open Knowledge Format (OKF) case study will follow once our early implementations complete, so you can see precisely how we define and measure success. Second, it makes an open offer: become our first named case study on founder-partner terms.
A quick note for clarity: every figure in the worked example below is an illustrative placeholder. Nothing here describes a real client or a real result.
What is OKF, in one paragraph
The Open Knowledge Format is an open specification (v0.1) published by Google Cloud on 12 June 2026. An OKF bundle is simply a directory of UTF-8 Markdown files, each with YAML frontmatter, a required type field, and reserved files such as index.md and log.md. It formalises the curated, portable, agent-readable “LLM-wiki” pattern so that AI agents can read your operational knowledge reliably. We are an independent implementation agency. We are not affiliated with, partnered with, or endorsed by Google.
The case study structure we will use
Every case study we publish will follow the same six-part structure. We are sharing it now so you can hold us to it.
1. Context
Who the client is, what they do, the size and shape of the team, and where their knowledge lived before. For example: a B2B SaaS company with documentation spread across a wiki, a support inbox, and several long shared documents.
2. The knowledge problem
The specific, measurable pain. Not “our docs are messy” but something concrete: support agents and AI assistants give inconsistent answers because the same policy is written four different ways across four tools, with no single source of truth an agent can read.
3. What we did
The work, step by step: discovery and source inventory, defining bundle types, agreeing the frontmatter schema, drafting and human-reviewing each Markdown file, and version-controlling the bundle in git. We always document decisions so the trail is auditable.
4. The OKF bundle delivered
The concrete artefact handed over: the directory structure, the file type values used, how index.md and log.md were populated, and how the bundle maps to the client’s real knowledge domains. You own this outright. There is no lock-in.
5. Outcome and metrics
This is where we prove value, against a baseline taken before we started. The metrics we intend to track include:
- Coverage: share of priority knowledge topics represented as valid OKF files (illustrative placeholder target: 0 to 85 per cent).
- Consistency: number of conflicting or duplicate answers removed (illustrative placeholder: 23 conflicts resolved).
- Agent answer quality: change in correct-answer rate when an AI assistant is grounded on the bundle (illustrative placeholder: a noticeable lift, measured on a fixed question set).
- Maintenance load: time to update a policy across the knowledge base after migration (illustrative placeholder: from hours to minutes).
Each real case study will state how the baseline was captured and who verified it. If we cannot measure it honestly, we will not claim it.
6. What the client can now do
The forward-looking payoff: ground internal AI agents and assistants on a clean, portable bundle; onboard new staff against a single source of truth; and extend the bundle themselves because it is plain Markdown in their own repository.
Be our first named case study
Because we are building our track record, we are offering founder-partner terms to a small number of early clients. In exchange for permission to document the work as a named, public case study (with your approval over every word and figure before anything is published), early partners receive preferential founder pricing and direct, hands-on involvement from the people doing the work.
This is a genuine exchange, not a discount gimmick. You get a spec-faithful OKF implementation at early-partner rates. We get a real, honest, verifiable case study built on your actual results. Nothing gets published without your sign-off, and any metric we cite will be one you can confirm.
Next step
The right place to start is a readiness audit. We assess your current knowledge sources, identify what is and is not agent-ready, and give you a clear scope for an implementation. From there you can decide whether founder-partner terms make sense for you.
Book an OKF Readiness Audit or see how our packages fit together.