A KnowledgeSource is a source from which signals of knowledge can be retrieved. It's a source from where data could be extracted to provide you with a KnowledgeMap of who knows what in your company.

A KnowledgeAsset is a discrete area of knowledge or expertise within your organisation — for example, a software module, a business process, a product line, or a regulatory domain. Assets are the subjects around which KnowledgeFabric builds your Knowledge Map: it tracks which KnowledgeHolders show signals of activity against each asset, so you can identify concentration risk at a glance.

A KnowledgeHolder is a person in your organisation who contributes knowledge signals — typically an employee, contractor, or collaborator identified through a connected KnowledgeSource (e.g. a Git committer, a Jira assignee, or a calendar meeting participant). KnowledgeFabric maps the knowledge each holder contributes to each asset and uses that data to calculate individual risk and resilience scores.

The Resilience Score reflects how well knowledge is distributed across your organisation for a given KnowledgeAsset or the overall manifest. A high score means multiple KnowledgeHolders share knowledge of an area, reducing dependency on any one person. A low score signals critical knowledge concentration — a key risk indicator. Scores are recalculated on every analysis run using graph algorithms (PageRank and Node Similarity) applied to the Knowledge Map.

When you trigger Analyse from the KnowledgeManifest screen, KnowledgeFabric runs the following steps:

1. Runs a health check on all connected KnowledgeSources.
2. Extracts knowledge signals from each source (Git commits, Jira issues, Confluence pages, etc.).
3. Maps signal identifiers to KnowledgeHolders in the database.
4. Applies time-decay weighting — older signals carry less weight than recent ones.
5. Writes weighted relationships into the Knowledge Map graph.
6. Recalculates resilience scores and risk metrics across all reports.

You can monitor progress in real time from the KnowledgeManifest dashboard.

KnowledgeFabric applies a time-decay model to all knowledge signals. A contribution from three years ago counts for far less than one from last month. The decay follows an exponential curve, and each integration type (Git, Jira, Confluence, etc.) has its own configured half-life. This ensures your Knowledge Map stays current and reflects who is actively contributing today, not just historically.

AI Asset Discovery uses a large language model to automatically suggest KnowledgeAssets based on your connected sources. For a Git repository, it analyses a compressed summary of the codebase to identify logical knowledge domains. Suggestions appear as "Discovered by AI" proposals that you can review, accept, or discard before they are added to your manifest. This feature requires a configured AI provider (Azure OpenAI or Ollama) set in your environment settings under Administration → AI Manager.

If your reports appear empty or unchanged after analysis, work through this checklist:

• At least one KnowledgeSource is connected and its health check passes (Configuration → Knowledge Sources).
• KnowledgeHolders have been created and their aliases match the identifiers used in your sources (e.g. Git email addresses, Jira usernames).
• KnowledgeAssets have at least one Signal Definition configured (Configuration → Knowledge Assets → Manage Signals).
• The analysis completed without errors — check the KnowledgeManifest progress log for any failed steps.

If the issue persists, download the support logs and contact us at support@knowledgefabric.io.

For the fastest resolution, follow these steps:

1. Go to Support → Support Logs and click Download Support Logs to save the latest application log file.
2. Go to Support → Submit Feedback, select Bug as the feedback type, describe the issue and the steps to reproduce it, and attach the downloaded log file.
3. Alternatively, email support@knowledgefabric.io or call (+356) 7925 1928 and share the log file with us.

Always include the log file when contacting support — it significantly speeds up diagnosis.