Skip to main content
Document support allows customers to upload files during conversations, enabling your AI agent to read and extract information from their documents for personalized, context-aware support.

​
Why Document Support Matters

File upload capabilities eliminate friction and enable smarter assistance: Eliminate Manual Data Entry - Customers share invoices, receipts, and specifications without manually entering details, reducing errors and saving time. Context-Aware Support - The AI can provide assistance based on actual document content rather than requiring customers to describe or transcribe information. Streamlined Workflows - Documents shared within conversation threads eliminate separate email submissions or portal uploads. Verification and Proof - Files serve as proof for refunds, warranty claims, identity verification, and compliance processes. Technical Troubleshooting - Customers can attach error messages, system logs, and technical documentation for targeted problem-solving.
This feature targets end customers sharing personal documents during support conversations, not businesses uploading knowledge bases to the chatbot system.

​
Supported File Types

botBrains accepts multiple document formats: PDFs (.pdf)
  • Invoices and receipts
  • Product manuals and specifications
  • Legal documents and contracts
  • Forms and applications
Office Documents
  • Word documents (.docx)
  • PowerPoint presentations (.pptx)
  • OpenDocument text (.odt)
  • OpenDocument presentations (.odp)
Text Files
  • Plain text (.txt)
  • Rich text format (.rtf)
  • Markdown (.md)
Data Formats
  • Spreadsheets (.xlsx, .csv)
  • JSON data (.json)
  • XML files (.xml)

​
How Document Upload Works

  1. Customer uploads - User clicks attachment icon and selects file from device
  2. Security scanning - File is scanned for malware and viruses
  3. Content extraction - AI reads and extracts relevant information from the document
  4. Context processing - AI understands document content and conversation context
  5. Personalized response - AI provides assistance based on actual file contents

​
Use Cases

​
Customer Support

Order and Purchase Verification: 
Customer: [Uploads invoice PDF] "I need to return this item.
When will I receive my refund?"

AI: [Reads invoice] "I can see from your invoice that you purchased
the wireless headphones on March 15th for $79.99. According to our
return policy, refunds are processed within 5-7 business days..."
Product Specifications: 
Customer: [Uploads product spec sheet] "Is this compatible with
the system I already have?"

AI: [Reads specifications] "Based on the spec sheet you shared,
this unit requires 220V power and has Bluetooth 5.0 connectivity.
Let me check compatibility with your existing setup..."

​
Technical Support

Error Messages and Logs: 
Customer: [Uploads error log .txt file] "The system keeps crashing.
Here's the error log."

AI: [Analyzes log] "I've reviewed your error log. The repeated
'Memory allocation failed' message at line 47 indicates a RAM issue.
Let's try these troubleshooting steps..."
Configuration Files: 
Customer: [Uploads config.json] "I can't get this to work. Here's
my configuration file."

AI: [Reviews configuration] "I found the issue in your configuration.
On line 12, the API endpoint is pointing to the staging server instead
of production. Change 'api.staging.example.com' to..."

​
Financial and Administrative

Receipt Processing: 
Customer: [Uploads receipt photo or PDF] "I need to claim warranty
on this purchase."

AI: [Extracts receipt data] "I can see from your receipt that you
purchased this on January 10, 2025 from our downtown store. Your
warranty is valid until January 10, 2027..."
Form Assistance: 
Customer: [Uploads partially filled form] "I'm not sure how to
complete section 3 of this form."

AI: [Reviews form] "Looking at section 3 of your form, you'll need
to provide your account number and transaction date. The account
number can be found on..."

​
Document Verification

Identity Verification:
  • Driver’s license uploads for account verification
  • Proof of address documents
  • Student ID for discount verification
  • Insurance cards for claims
Proof of Purchase:
  • Receipts for returns and exchanges
  • Invoices for warranty claims
  • Order confirmations for shipping issues
  • Payment confirmations for disputes

​
Configuration

Enable document upload for your deployment:
  1. Go to Deploy → [Your Deployment] → Settings
  2. Enable Document Upload
  3. Configure accepted file types (or use all supported types)
  4. Set maximum file size limits
  5. Review security settings
  6. Deploy to production

​
Guidance Considerations

Train your AI to handle document-based interactions: 
"When customers upload documents, carefully read and extract relevant
information. Reference specific details from the document in your
response to show you've understood the content."

"If a document is unclear, missing information, or in an unexpected
format, politely ask the customer for clarification or request an
alternative document format."

"For documents containing sensitive information (financial data,
personal details), acknowledge receipt and handle the data according
to privacy guidelines. Don't repeat sensitive information unnecessarily."

​
Best Practices

​
Clear Upload Instructions

Guide users on what documents to share: 
"Click the attachment icon to upload your invoice or receipt"
"Accepted formats: PDF, Word documents, images, and text files"
"Maximum file size: 10MB"

​
Document Quality Guidance

Help users provide usable documents: 
"For best results, ensure text in PDFs is selectable (not scanned images)"
"If uploading a photo of a document, ensure it's well-lit and in focus"
"For multi-page documents, upload the complete file rather than individual pages"

​
Privacy and Security Communication

Be transparent about document handling: 
"Your document will be securely processed and used only to assist
with your current request"
"We scan all uploads for security before processing"
"Documents are retained according to our privacy policy"

​
Fallback Handling

Prepare for unreadable documents: 
Guidance: "If you cannot extract information from an uploaded
document, let the customer know and ask them to either:
1. Provide the key information in text form
2. Upload a different version of the document
3. Upload a clearer image or scan"

​
Technical Considerations

​
Security and Safety

Malware Protection:
  • All uploads scanned for viruses and malware before processing
  • Infected files blocked and not processed
  • User notified if file cannot be processed
Content Moderation:
  • Currently implementing content moderation guardrails
  • Protection against inappropriate content
  • Team protection from harmful material
  • Reporting mechanisms for policy violations
Document content moderation is still being enhanced. botBrains is working to implement the same level of content moderation guardrails for documents as currently exists for audio and image files.

​
Privacy and Data Protection

Data Handling:
  • Documents processed for current conversation context
  • Retention policies configurable per organization
  • GDPR compliance for personal data
  • Encryption in transit and at rest
User Control:
  • Customers can see which documents were shared
  • Option to request document deletion
  • Clear privacy notices in deployment

​
File Size and Limits

Recommended Limits:
  • Maximum file size: 10MB (configurable)
  • Files larger than limit rejected with clear error
  • Multiple files can be uploaded per conversation
  • Pagination for multi-page documents
Performance:
  • Processing time varies by file size and type
  • Typical processing: 2-5 seconds for standard documents
  • Large or complex files may take longer
  • Progress indicators shown during upload

​
Format Compatibility

Text Extraction:
  • PDFs with selectable text work best
  • Scanned PDFs (images) have lower accuracy
  • Office formats (.docx, .xlsx) fully supported
  • Plain text files processed instantly
Complex Formatting:
  • Tables and structured data supported
  • Charts and graphs in documents (limited understanding)
  • Multi-column layouts handled
  • Embedded images processed where relevant

​
Frequently Asked Questions

Possible causes:
  1. File size exceeds limit
  2. Unsupported file type
  3. Internet connection interrupted
  4. Browser compatibility issue
  5. File is corrupted
Solutions:
  • Check file size (must be under limit)
  • Verify file type is supported
  • Try compressing large PDFs
  • Check internet connection stability
  • Try different browser
  • Re-save document and try again
Possible causes:
  1. Scanned PDF (image-based, not text)
  2. Corrupted file
  3. Password-protected document
  4. Unsupported language
  5. Poor image quality (for scanned documents)
Solutions:
  • Use OCR to convert scanned PDFs to text
  • Remove password protection
  • Re-save document in supported format
  • Provide key information in text message
  • Upload clearer scan or photo
If AI unnecessarily exposes sensitive data:Solutions:
  • Update guidance to handle sensitive information carefully
  • Train AI to reference documents without repeating sensitive details
  • Review conversation and provide feedback
  • Contact support if issue persists
Possible causes:
  1. Complex document layout
  2. Poor quality scan
  3. Ambiguous content
  4. AI misinterpretation
Solutions:
  • Provide clarifying information in text
  • Upload clearer version of document
  • Point out specific sections AI should focus on
  • Use simpler document format (e.g., .txt instead of complex .pdf)

​
Example Scenarios

​
Return Processing

Customer scenario: Needs to return defective product Action: Customer uploads purchase receipt (PDF) AI reads:
  • Purchase date: March 15, 2025
  • Item: Wireless Headphones Model XYZ
  • Price: $79.99
  • Store: Downtown Location
AI response: “I can see from your receipt that you purchased the Wireless Headphones Model XYZ on March 15, 2025 for $79.99 from our downtown store. Since you’re within the 30-day return window, you’re eligible for a full refund. Would you like to initiate a return shipment or return in-store?”

​
Technical Troubleshooting

Customer scenario: Application configuration issue Action: Customer uploads config.json file AI reads: 
{
  "api_endpoint": "api.staging.example.com",
  "timeout": 5000,
  "retry_attempts": 3
}
AI identifies: API endpoint pointing to staging instead of production AI response: “I found the issue in your configuration file. On line 2, your API endpoint is set to ‘api.staging.example.com’, which is the staging server. For production use, please change this to ‘api.example.com’. This should resolve your connectivity issues.”

​
Warranty Claim

Customer scenario: Claiming warranty on broken appliance Action: Customer uploads original invoice and photos of defect AI reads invoice:
  • Purchase date: January 10, 2025
  • Product: Coffee Maker Pro 3000
  • Warranty: 2 years
AI combines: Invoice data + visual defect assessment + warranty policy AI response: “Based on your invoice from January 10, 2025, your Coffee Maker Pro 3000 is covered under our 2-year warranty until January 2027. I can see the cracked water reservoir in your photos, which is covered. I’ll help you start a warranty claim…”

​
Next Steps

Now that you understand document support:
  • Vision - Enable image understanding for visual documents
  • Audio - Allow customers to describe issues with voice messages
  • Data Export - Export conversations including uploaded documents
  • Deploy - Configure document upload in your deployments
Document upload capabilities eliminate friction and enable smarter, context-aware customer support. Enable file sharing strategically based on your support workflows and customer needs.