Skip to main content

Knowledge Assistant — Internal

Knowledge Assistant (Internal) is a variant of a RAG system for the organization's internal users. It helps employees find and use corporate knowledge: policies, procedures, regulations, and technical documentation.

Purpose​

Employees spend significant time searching for information across internal systems. The Knowledge Assistant makes it possible to:

  • ask a question in natural language
  • receive an answer with links to sources
  • reduce the time spent searching for information

Use Cases​

Typical scenarios for an internal Knowledge Assistant:

  • HR policies — questions about leave, benefits, and onboarding procedures
  • IT support — setup instructions, FAQs about internal systems
  • Regulatory requirements — searching regulatory documents and standards
  • Onboarding — helping new employees orient themselves to processes
  • Internal documentation — searching technical and business documents
  • Legal questions — routine questions about contracts and policies

Data Sources​

The Knowledge Assistant connects to the following sources:

  • Confluence, SharePoint, internal wikis
  • regulatory and normative documents
  • internal policies and procedures
  • technical specifications and manuals
  • IT support knowledge bases
  • corporate orders and regulations

Key Challenges​

When implementing an internal Knowledge Assistant, the following difficulties arise:

  • Access control — different employees have access to different documents; the system must take roles and departments into account
  • Document freshness — outdated documents lead to incorrect answers; an index update process is required
  • Multiple formats — documents are stored in various formats (PDF, Word, PowerPoint, HTML); each format requires parsing
  • Answer quality — answers must be accurate; errors undermine user trust
  • Duplicates and versioning — the same information may exist in several documents of different versions

Metrics​

The effectiveness of an internal Knowledge Assistant is measured by the following metrics:

  • Query volume — number of questions per day/week
  • Answer accuracy — percentage of correct answers (assessed by experts or users)
  • Time savings — reduction in time spent searching for information compared to manual search
  • User satisfaction — quality assessment through surveys and feedback
  • Coverage — percentage of questions the system was able to answer
  • Adoption rate — share of employees regularly using the system