Skip to main contentFile Upload
The File Upload section under the Knowledge Base module allows Admins to seamlessly upload and manage multiple documents
that can be leveraged within Axoma’s intelligent chat interface for contextual question answering and information retrieval.
Key Features
Bulk Upload Support
Users can upload multiple files simultaneously, making it efficient to populate the knowledge base with relevant enterprise
documents such as PDFs, DOCX files, and more.
Access Rights Assignment
During upload, users can assign predefined Access Rights
to each file. This ensures that only the authorized Groups or roles
within the organization can access or query specific documents.
Tag Association
Tags created under the Tags section can be assigned to each file
to facilitate organized classification, sorting, and retrieval
of documents based on business context, departments, or use cases.
Parser Preferences
When uploading documents to the Knowledge Base, Parser Preferences allow you to choose how the file content is processed for optimal retrieval and accuracy. Three parser modes are available:
Quick - Ideal for clean, text-only documents. It skips image and table processing to ensure fast parsing using basic text chunking (chunk_size: 1024, overlap: 200 by default).
By clicking on the “Customize” button within this parser setting, admins or superadmins can modify the chunk size and overlap values to better suit specific document structures or use cases.
Smart - Recommended for standard documents containing light imagery or tables. It performs OCR and detects basic structures like headings and images for balanced performance and accuracy.
Ultra - Designed for complex, image-heavy files. It applies advanced OCR, image scaling, and deep structure recognition to maximize extraction precision.
Choose the parser mode that best suits your document type to improve knowledge base performance.
Users can select the appropriate parser based on the nature and complexity of the document being uploaded.
Chat Integration
Once a document is uploaded and processed, it becomes accessible within Axoma’s chat interface (DocuChat). Users can interact with the content through natural language questions, allowing AI-driven document interrogation and intelligent retrieval of relevant information.
This feature empowers organizations to centralize unstructured knowledge, control access, and enable AI-powered insights all within a governed, secure environment.
Once all chat integrations are successfully configured, Admins and Super Admins can run and test the app in Draft
mode to ensure everything functions as expected before publishing. For detailed configuration and access controls,
refer to the App Settings section. 
