Enterprise Knowledge gives AI business understanding — not just more information.
It is an enabling pillar of AI Fabrix. Enterprise Knowledge supplies business context; Operational Trust applies boundaries; Role Assistants perform governed work; Evidence Fabrix proves the outcome.
Describe the business once. Reuse entities, metadata, and relationships across Role Assistants. Existing applications remain the systems of record.
Enterprise Knowledge is not only a data catalog or search index. It is the business model AI uses to understand what exists, how it connects, what it means, who may use it, and what work can be requested.
Give AI understanding — not only retrieval
Enterprise information does not have equal meaning or authority. A regulation is different from a company policy. A current agreement is different from an obsolete document. Customer, contract, project, and product information may be related even when they live in different systems.
Traditional RAG can retrieve semantically similar content. That alone does not tell AI which business context applies, which source is authoritative, which version is current, or which information the user may receive.
Smart RAG with business metadata
Smart RAG combines semantic relevance, business metadata, and permission-aware filtering.
Metadata can describe the active customer, contract, project, or case; source authority; relationships; classification; ownership; permissions; version and validity. Role Assistants use this metadata to narrow retrieval to the active role and outcome before content is supplied to AI.
At retrieval time, Operational Trust limits which Enterprise Knowledge capabilities and content are available for the active user and case. Relevant is not the same as permitted — see Dimensions and protection.
Smart RAG is one Enterprise Knowledge capability. It complements structured business data and governed operations; it does not replace them. Practice detail: Records search and AI context.
Viewpoints: certified business pictures
Some work needs a predefined business picture rather than more documents or a list of individual records.
A Viewpoint gives a Role Assistant a certified, read-only perspective — for example revenue by customer, pipeline by stage, or actual cost by cost centre. It is calculated when used under Operational Trust, while systems of record remain authoritative. It is not a second reporting warehouse.
Smart RAG retrieves permitted content; a Viewpoint provides an approved business picture for the current decision. See Viewpoints and joined business context.
Enterprise Knowledge helps AI retrieve the right information — and understand why it applies.
What Enterprise Knowledge owns (and does not)
| Concern | Owner |
|---|---|
| Business meaning, metadata, relationships, governed search context, sync freshness for connected systems | Enterprise Knowledge |
| Current authorized business facts (“what is true now”) | Enterprise Reality |
| Certified reusable business proof | Evidence Fabrix |
| Policy, identity, dimensions, certification gates, capability admission | Operational Trust |
| Decision Engine orchestration and continuation | Architecture / Runtime |
Source systems remain authoritative for live records. Enterprise Knowledge organizes and exposes governed business context from certified integrations — it does not replace Reality, certify Evidence, or execute Role Assistant work.
Organization documents and uploads usually enter as document storage / Knowledge first. Approving document metadata for search is not the same as certifying Evidence. See Evidence Fabrix and Document approval and governance.
Why it matters
AI cannot perform useful enterprise work when it sees disconnected system objects. Traditional platforms expose CRM accounts, file paths, and ticket IDs. AI Fabrix creates business meaning:
Customer
├── contacts
├── deals
├── contracts
├── documents
├── support cases
├── owners
└── available work
That is the difference between connecting systems and giving AI business understanding. Prefer a minimal set of business entities and metadata — quality over quantity — then extend when the business needs a new entity. See Business entities and resource types.
Business metadata
Business metadata explains what data means in business language — not only how a vendor stores it. Resource types (customer, contract, document, task) define the business entity before system structure. Detail: Business metadata.
Cross-system relationships
Cross-system relationships link business entities across systems so no single application has to hold the full story. Detail: Cross-system relationships.
How it works
Systems of record (authoritative)
↓
Business metadata and resource types
↓
Cross-system relationships
↓
Business context (Enterprise Knowledge)
↓
Smart RAG / Viewpoints / governed capabilities
↓
Operational Trust at use (permitted for this person and case)
Datasources sync external records into governed business entities. Dimensions and protection from Operational Trust apply so context stays within policy. Uncertified or out-of-scope data does not become assistant context.
When several Connected Systems share one business object, write and list authority declares preferred mutations and authoritative lists from metadata — without custom multi-write glue.
Governed capabilities from business context
Enterprise Knowledge defines what AI can request through governed business capabilities — not raw API calls. AI requests the capability; AI Fabrix handles system-specific execution under Operational Trust. Admission is fail-closed. See Governed capabilities.
Enterprise MCP
Enterprise MCP is the governed AI interface to Enterprise Knowledge — business language, not raw vendor APIs. It is compiled after publish and certification — not a separate product beside Role Assistants. Detail: Enterprise MCP.
Example
A customer review may require CRM accounts and deals, signed contracts, open support cases, project documents, and owner information. Enterprise Knowledge assembles that into one business context; Smart RAG and Viewpoints supply the right shape of knowledge; Operational Trust keeps it within role and dimension scope before a Role Assistant acts.
Business value
Enterprise Knowledge helps AI move from searching systems to understanding the business, connect cross-system context safely, reuse definitions across Role Assistants, and keep systems of record authoritative.
One-line summary
Enterprise Knowledge helps AI retrieve the right information — and understand why it applies — under Operational Trust.