What is Knowledge Base in Workforce?

Knowledge Base in Workforce is your Teammate's information repository. It stores documents, images, spreadsheets, and other data types in a vector store, enabling your AI agent to perform in-depth analysis, execute intelligent searches, and provide contextually relevant responses.

Think of it as: Your Teammate's reference library that it can search through and understand to answer questions accurately.

What can Knowledge Base do?

Knowledge Base empowers your Teammate to:

• Answer questions based on your uploaded documents
• Summarize content from contracts, reports, or manuals
• Perform analysis on spreadsheets and structured data
• Execute calculations and data transformations using Code Interpreter
• Extract insights from presentations and visual content
• Generate new files (CSVs, reports, charts) from uploaded data
• Reference precise details to provide accurate, context-aware responses

Best for:

• Customer support agents that need access to product manuals
• Financial analysis using company reports and spreadsheets with calculations
• HR teammates answering policy questions from employee handbooks
• Sales agents referencing product specifications and pricing sheets
• Data analysis requiring programmatic manipulation of uploaded files
• Research tasks requiring access to technical documentation

Key benefits of using Knowledge Base

• Enhanced Contextual Understanding: Your Teammate understands the specific context of your business, not just general information.
• Improved Accuracy: Teammates reference precise details from your documents rather than relying on general knowledge.
• Computational Capabilities: With Code Interpreter enabled, perform calculations, data transformations, and generate new files from uploaded content.
• Scalability: Upload large volumes of documents to create a comprehensive information base.
• Low-Code Integration: Simply upload files without special formatting or technical configuration.
• Multi-Document Support: Use multiple knowledge files within a single Teammate workflow.

What types of data can I upload?

Workforce supports a wide variety of data types:

Text Documents

• Word documents (.docx, .doc)
• PDFs (.pdf)
• Plain text files (.txt)
• Markdown files (.md)
• HTML files (.html, .htm, .xhtml)

Spreadsheets

• Excel files (.xlsx)
• CSV files (.csv)

Programming & Scripts

• Python (.py)
• JavaScript (.js)
• TypeScript (.ts)
• Java (.java)
• C/C++ (.c, .cpp, .cs)
• Ruby (.rb)
• PHP (.php)
• Shell scripts (.sh)
• CSS (.css)

Images

• JPEG (.jpg, .jpeg)
• PNG (.png)
• GIF (.gif)

Presentations

• PowerPoint slides (.pptx)

Other Formats

• JSON (.json, .ldjson)
• XML (.xml)
• iCalendar (.ics)
• LaTeX (.tex)
• ZIP archives (.zip)
• TAR archives (.tar)
• DEC files (.dec)

Analysis capabilities for uploaded documents

When you upload documents, you can enable different analysis types depending on your needs:

Text Analysis (Vector Search)

What it does: Enables semantic search across your document content

Best for:

• Reading and searching document text
• Finding relevant sections based on questions
• Understanding document context and meaning

Use cases:

• Product manuals for customer support
• Policy documents for compliance questions
• Technical documentation for troubleshooting

Code Analysis (Code Interpreter Tool)

What it does: Allows the Teammate to generate and execute Python code to interact with your files

Best for:

• Performing calculations on spreadsheet data
• Transforming and manipulating data
• Creating new files (CSVs, reports, visualizations)
• Extracting structured information from documents

Use cases:

• "Calculate the average sales over the past year" from an Excel file
• "Create a CSV of all customers with revenue over $10,000" from a spreadsheet
• "Based on this PDF, create a list of all people mentioned with their titles"
• Generate charts and visualizations from data

What you'll see: When Code Interpreter runs, you'll see:

• The Python code that was generated
• Execution output and results
• Download links for any files created
• Debug information if errors occur

Image Analysis (AI Vision)

What it does: Processes visual content to understand images, diagrams, and charts

Best for:

• Extracting information from diagrams
• Understanding charts and graphs
• Reading scanned documents
• Analyzing photographs

Use cases:

• Process flow diagrams
• Organizational charts
• Scanned receipts or forms
• Product images

Using multiple Knowledge files

Teammates are highly modular and can work with multiple Knowledge files simultaneously.

Example scenario: Upload both technical manuals AND financial spreadsheets, and your Teammate will seamlessly answer questions or perform analysis across both document sets.

Common combinations:

• Product manuals + pricing sheets (with calculations)
• Employee handbook + benefits documentation
• Technical specs + compliance guidelines
• Training materials + company policies
• Multiple spreadsheets for comprehensive data analysis

Updated on: 03/11/2025