Hi, I am working with a construction company that generates feasibility reports for land parcels across the United States. Currently, we use ChatGPT along with publicly available data sources, including: County / City Public Records Portals GIS Maps Zoning Ordinance Documents Flood & Wetland Data Sources Utility Availability Maps Google Maps / Satellite Images We have already built a basic AI agent, but it is only ~70% accurate, and the image extraction + report generation is not working properly. We are looking for an experienced freelancer to improve or rebuild the entire AI-based feasibility system. Scope of Work: The system should: 1. Input Handling Accept US property address (even incomplete or unformatted) Automatically detect: City County Jurisdiction 2. Data Extraction Fetch and structure data from relevant public sources: Property Details Owner Name Parcel Number Lot Size (acres / sqft) Legal Description Zoning Code Property Status Zoning & Setbacks Pull from city ordinance sources Include: Setbacks (front, rear, side) Height restrictions Allowed development types 3. Utilities & Infrastructure Electricity availability + provider Water availability (city / well) Sewer / Septic status Utility connection fees (if available) 4. GIS & Environmental Data Flood Zone Wetland Status Protected zones (if applicable) Sewer / septic map overlays Nearby comps / neighborhood insights 5. Image Capture (CRITICAL REQUIREMENT) This is our main problem area. The system must: Capture relevant images from: GIS maps Google Maps / Satellite Public portals Ensure images are: Correct for the property High quality Properly labeled Embed images into the final report correctly 6. Report Generation Generate a professional report (DOCX / PDF) Structured sections like: Lot Details Zoning Utilities GIS Analysis Development Potential Clean formatting with images embedded (not broken links) 7. Web Application (Required) We want this system to run on our company website: User Login System Address Input Form Report Generation Dashboard Download Report (PDF/DOCX) Payment Integration (Pay-per-report model) Current Issue (Important): Existing system gives partial correct data Images are incorrect or not embedded properly Output format is inconsistent (HTML not converting cleanly to report)