Agentic RAG & AI Document Agents
The elDoc File Management Module includes Agentic RAG, enabling users to interact with enterprise documents and knowledge directly through AI Chat. Click AI Chat in the File Management Module to get started. AI Chat is more than a conversational interface. It combines Agentic RAG with AI Document Agents, allowing users to ask questions in natural language, retrieve information from authorized enterprise documents, and instruct AI Agents to perform document-related tasks.
With Agentic RAG and AI Agents, users can:
| AI Document Agent Capability | What It Can Do |
|---|---|
| Extract Data | Identify and extract required information from documents. |
| Classify Documents | Analyze content and classify documents according to provided instructions. |
| Perform Deep Research | Research information across multiple enterprise documents and sources available to the Agent. |
| Rename Files | Analyze document content and propose appropriate file names. |
| Reorganize Documents | Classify, group, move, and organize documents into appropriate folder structures. |
| Create & Edit Documents | Create new documents or modify existing content based on instructions. |
| Complete Data | Complete document information based on provided instructions and available context. |
| Validate & Verify Information | Check information across documents and identify discrepancies. |
| Analyze Compliance | Review documents against provided requirements, policies, or criteria. |
| Identify Risks & Inconsistencies | Detect potential issues, inconsistencies, or missing information across documents. |
| Summarize Documents | Generate summaries of individual or multiple documents. |
| Search Enterprise Knowledge | Search and retrieve relevant information across authorized enterprise content. |
Access & Permissions
AI Agents operate strictly within the AI File Management environment and respect existing file and folder permissions. An AI Agent can access and work only with documents, folders, and enterprise knowledge that the user is authorized to access. Existing access controls therefore continue to apply when using Agentic RAG and AI Agents.
Selecting an LLM Model
Depending on the organization’s GenAI configuration, AI Chat may provide access to different LLM models. The selected model can affect processing speed, reasoning capabilities, performance, and accuracy. This allows the appropriate model to be selected depending on the complexity and requirements of the task.
AI Chat & Task History
AI Chat conversations and AI Agent tasks are automatically saved in the history. Users can return to previous interactions to review earlier requests, results, and AI tasks.
Tip: Give AI Agents Specific Instructions. For better results, prompts should clearly describe what needs to be done, which documents should be analyzed, the criteria to apply, and the expected result.
Instead of: Organize these documents. Use a more specific prompt: Analyze the selected documents, classify them by document type and year, rename each file using Document Type_Year, and organize them into the corresponding folders.
Important: For scanned or image-based documents, make sure the OCR Text Index status is Ready before asking an AI Agent to analyze their content. When GenAI is enabled, OCR and text indexing are performed automatically during file upload; no separate OCR step is required.