I’m currently migrating our workflows from ArcGIS Online to ArcGIS Enterprise and am working on improving how we manage easement/permit documentation tied to parcel features.
Right now, I’m beginning to link scanned easement documents (PDFs) to parcel features (either via attachments or related tables). The next step I’d like to tackle is automating attribute population from those scanned documents instead of manually keying everything in.
Ideal workflow:
Batch associate scanned permit/easement documents to parcel features
Extract key fields from each document (Grantor, Grantee, Legal Description, Date, Parcel ID, etc.)
Auto-populate corresponding GIS attributes
Preferably using some kind of template-based or semi-structured document extraction
Questions:
Has anyone built a workflow like this within the Esri ecosystem?
Are you using tools like ArcGIS Pro + ArcGIS Notebooks, or integrating external OCR/AI tools?
Any experience using:
FME for document parsing?
AI/OCR tools (e.g., Azure Form Recognizer, AWS Textract, etc.) tied back into feature services?
How well do template-based extractors handle variation in scanned legal documents?
Any recommended architecture for scaling this (thousands of documents)?
I’m especially interested in repeatable, batch-friendly workflows that minimize manual QA but still maintain accuracy for legal records.
Appreciate any insight, examples, or lessons learned- especially if you’ve implemented this in a utility, ROW, or parcel management context.
