Claims processing · Production
Nanonets automates insurance claim form processing for Defined Physical Therapy Network
The problem
Defined Physical Therapy Network manually entered data from scanned insurance claim forms into their CRM, taking about 11 minutes per patient, while the team processed over 60,000 pages annually.
Workflow diagram · grounded in source
1
Forms ingested from source
trigger
“Nanonets ingests the incoming insurance claim forms directly at the source, which is Gmail or G-Drive”
2
OCR data extraction
ai_action
“The AI-powered custom OCR model at Nanonets extracts the relevant data points from the claim forms”
3
Post-processing data cleanup
validation
“This data then flows through a number of post-processing formatting steps such as finding and replacing characters, removing characters, etc. to clean up the extracted data”
4
Export to CRM via webhooks
integration
“This data is then transferred to webhooks where it gets exported to the internal CRM at Defined Physical Therapy Network”
5
Claim categorised and filed
output
“The claim gets categorised and the team can file the claim with the respective payer”
Reported outcome
Nanonets automated the end-to-end manual process, ingesting forms from Gmail or G-Drive, extracting data with an AI-powered OCR model, and exporting cleaned data to the internal CRM via webhooks so claims can be categorised and filed with the respective payer.
Reported metrics
Manual filing time per patient11 minutes
Annual pages processed manuallyover 60,000 pages
Reported stack
NanonetsOCRGmailG-DriveCRM
Source
https://nanonets.com/customer-success-story/defined-physical-therapy-automates-insurance-claim-form-processing
Read source ↗Frequently asked questions
What did this team achieve with this AI workflow?
Nanonets automated the end-to-end manual process, ingesting forms from Gmail or G-Drive, extracting data with an AI-powered OCR model, and exporting cleaned data to the internal CRM via webhooks so claims can be categ…
What tools did this team use?
Nanonets, OCR, Gmail, G-Drive, CRM.
What results were reported?
Manual filing time per patient: 11 minutes; Annual pages processed manually: over 60,000 pages (source-reported, not independently verified).
How is this claims processing AI workflow structured?
Forms ingested from source → OCR data extraction → Post-processing data cleanup → Export to CRM via webhooks → Claim categorised and filed.