Your knowledge is scattered, tool-locked, and hostile to agents
Most organisations hold their operational knowledge in a dozen disconnected places: a wiki nobody trusts, a data catalog only analysts open, support macros buried in a helpdesk, onboarding docs in shared drives, and tacit process knowledge that lives only in people’s heads. Each store has its own format, its own access model, and its own decay rate.
That arrangement was tolerable when only humans read it. It is a liability now that AI agents are expected to act on it. An agent cannot reason over a permission-walled wiki, a proprietary catalog API, or a PDF export. It needs curated, portable, version-controlled, plain-text knowledge it can read directly. That is exactly what Google’s Open Knowledge Format (OKF) standardises.
The OKF Readiness Audit is the low-risk first step. It is a fixed-fee engagement that tells you, honestly and concretely, how far your current knowledge is from being agent-ready, and what it would cost to close the gap.
Who this is for
- Operations, data, and platform leaders who want their internal agents to give correct, current answers.
- Teams piloting AI assistants that keep hallucinating because the underlying knowledge is unstructured or stale.
- Knowledge and documentation managers asked to make their content “AI-ready” without a clear definition of what that means.
- Engineering leaders evaluating OKF as a standard before committing build resource.
What the audit assesses
We work against the OKF v0.1 specification (open spec, Google Cloud, 12 June 2026). A conformant OKF bundle is a directory of UTF-8 Markdown files with YAML frontmatter, a required type field, recommended title, description, resource, tags, and timestamp fields, and reserved index.md and log.md files. We measure your current state against that target.
| Assessment area | What we examine |
|---|---|
| Knowledge source inventory | Every system holding operational knowledge: wikis, data catalogs, docs, support KBs, drives, code comments, runbooks |
| Current agent-context setup | How agents are fed today: RAG pipelines, pasted context, vector stores, prompt stuffing, or nothing |
| Content quality and freshness | Duplication, contradiction, staleness, ownership gaps, and tacit knowledge not yet written down |
| OKF conformance gaps | Distance from v0.1: file structure, frontmatter completeness, type usage, index.md and log.md readiness |
| Prioritised opportunity map | Which domains deliver the most agent value for the least conformance effort |
What you receive
| Deliverable | Description |
|---|---|
| Written readiness report | A plain-language assessment of your knowledge estate and its agent-readiness, suitable for sharing with leadership |
| Scored gap analysis | Each candidate domain scored against OKF v0.1 conformance criteria, with the specific gaps named |
| Recommended bundle architecture | A proposed directory and type structure for your priority domain, with a progressive-disclosure index.md outline |
| Prioritised roadmap | A sequenced plan: which domain to formalise first, second, third, and why |
| Fixed-fee implementation quote | A firm price to build the first bundle, with no open-ended hours |
The process and timeline
- Kickoff (day 1). A 60-minute session to confirm scope, name your knowledge sources, and grant read access where needed.
- Inventory and review (days 2 to 6). We catalogue your sources, sample content for quality and freshness, and document your current agent-context setup.
- Conformance scoring (days 6 to 9). We score candidate domains against OKF v0.1 and draft the recommended bundle architecture.
- Report and walkthrough (days 9 to 12). We deliver the written report and gap analysis, then walk you through the roadmap and fixed-fee quote on a call.
A typical engagement completes in 1 to 2 weeks.
What we need from you
- A named point of contact and 2 to 3 hours of stakeholder time across the engagement.
- Read access to the knowledge sources in scope, or representative exports.
- A short brief on where agents will consume the knowledge: internal copilots, customer-facing assistants, or a RAG pipeline.
The outcome
You finish with a clear, evidence-based picture of your knowledge estate, a defensible decision on what to formalise first, and a firm price to do it. No vague “digital transformation” narrative, just a scored gap analysis and a costed plan you can take to a budget holder.
How it leads into implementation
The audit is deliberately the step-in-the-door offer. The roadmap and architecture become the build brief for OKF implementation, where we turn your chosen domain into a conformant, maintained OKF bundle. If you migrate an existing knowledge base rather than author fresh, the migration sprint picks up from the same audit. The audit fee is credited against implementation if you proceed within 90 days.
Independence note: we are an independent implementation agency. OKF is an open Google Cloud specification. We are not affiliated with or endorsed by Google.
Ready to find out how agent-ready your knowledge is? Book a readiness audit.