What it is
A Unified Knowledge Base aggregates structured and unstructured content — articles, PDFs, CRM records, web pages — into one searchable layer that AI agents can query at runtime. It acts as the long-term memory and factual backbone for agent responses. Unlike traditional knowledge management systems, a UKB is optimized for semantic retrieval, not just keyword search.
Why it matters
Agents without a strong knowledge base hallucinate or give outdated answers, destroying customer trust. A well-maintained UKB dramatically improves response accuracy and reduces escalations. For mid-market companies, it also eliminates the costly silo problem where support, sales, and ops teams each maintain separate documentation.
Key components
- semantic indexing
- content connectors
- retrieval pipeline
- access controls
- freshness scoring
How it connects
Salesforce Data Cloud and Einstein feed directly into Agentforce's knowledge layer, allowing agents built in Agent Builder to query grounded, up-to-date company content before generating a response.
Good to know
Garbage in, garbage out — a UKB is only as good as its curation discipline. Assign a content owner and set quarterly review cycles before going live with agents.
Related terms
Data Cloud
Salesforce's unified data platform that centralizes customer data from all sources, making it available for agents, analytics, and insights in real time.
Grounding
Connecting AI responses to your actual business data so agents answer with facts, not hallucinations.
Knowledge Management
The systems and processes for organizing, storing, and making business information accessible to AI agents so they can provide accurate, contextual answers.
RAG (Retrieval-Augmented Generation)
A technique that makes AI smarter by fetching relevant information from your data before generating a response. The AI "looks it up" instead of guessing.