Step-by-step guide to uploading and extracting data from construction documents using Smthy's AI-powered processing.
Processing your first PDF document in Smthy is straightforward, but knowing the right steps will help you get the best results from our AI extraction system. This guide walks you through the entire process from upload to review.
What You'll Learn
• How to upload PDFs for optimal AI processing
• What information gets extracted automatically
• How to review and refine extracted data
• Tips for getting better results from complex documents
1
Prepare Your Document
Before uploading, ensure your PDF is optimised for AI processing.
✅ Best Document Types
• Landscape plans with plant schedules
• Supplier quotes and price lists
• Material specifications
• Construction details with quantities
⚠️ Quality Requirements
• Clear, readable text (not handwritten)
• High resolution (300 DPI minimum)
• File size under 50MB
• PDF format (not image files)
2
Upload Your Document
Upload your PDF through the project interface or dashboard.
Upload Methods
Navigate to your project or create a new one
Click the "Upload Documents" button
Drag and drop your PDF or click to browse
Add a descriptive name for the document
Select the document type (quote, plan, specification)
Pro Tip: Batch Uploads
You can upload multiple related documents at once. Smthy will process them in the background while you continue working.
3
AI Processing
Smthy's AI analyses your document and extracts structured data automatically.
What Gets Extracted
• Plant names and quantities
• Pot sizes and specifications
• Unit prices and totals
• Material descriptions
• Supplier information
• Project details
Processing Time
• Simple quotes: 30 seconds - 2 minutes
• Complex plans: 2 - 5 minutes
• Large documents (20MB+): 5 - 10 minutes
You'll receive a notification when processing is complete.
4
Review Extracted Data
Review the extracted information and make any necessary corrections.
Review Checklist
• Verify plant names are correct (botanical and common)
• Check quantities and pot sizes match the document
• Confirm pricing information is accurate
• Ensure supplier details are complete
• Add any missing categories or tags
Making Corrections
Click on any extracted item to edit it. Your corrections help improve AI accuracy for future documents.
• Edit text directly in the extracted data fields
• Add missing information manually
• Merge duplicate entries
• Delete incorrect extractions
Common Issues & Solutions
AI Missed Some Information
This is normal, especially for complex documents. You can:
• Add missing items manually
• Re-upload a clearer version of the document
• Use the "Re-process" option after making corrections
Incorrect Plant Names
AI sometimes misreads botanical names. Always double-check:
• Compare against the original document
• Use plant name databases for verification
• Add common names if missing
Processing Failed
If processing fails, try these solutions:
• Check file size is under 50MB
• Ensure PDF is text-based, not just images
• Try uploading during off-peak hours
• Contact support if issues persist
Getting the Best Results
Document Quality Tips
• Use the original PDF from suppliers, not photocopies
• Ensure text is selectable in the PDF
• Remove password protection before uploading
• Include complete plant schedules and price lists
Naming Conventions
Use descriptive names for better organisation:
• "Smith2024_NurseryQuote_AustralianNative.pdf"
• "Jones2024_IrrigationSpec_HunterSystems.pdf"
• "MelbUni2024_PlantSchedule_Phase1.pdf"
You've Succeeded When...
Success Indicators
• All plant names are correctly extracted and spelled
• Quantities and pot sizes match your original document
• Pricing information is accurate and complete
• Items are automatically categorised appropriately
• You can easily search and find the extracted items
What's Next?
Now that you've processed your first document, learn how to use semantic search to find exactly what you need.