SafeRide Health automates background verification for vendors saving 52% of their cost using Nanonets

1. Introduction: SafeRide Health provides non-emergency medical transportation to Medicare and Medicaid health plan members and care provider patients for vulnerable populations across the United States.
2. Headquarters: Texas, United States of America
3. Founded: 2016
4. Founders: Robbins Schrader, Ben Salter, and Whit Schrader
5. Industry: Healthcare
6. Employee Count:~250-501

Summary: Scaling the process of background verification of the network of vendors, drivers and vehicles and maintaining a database of relevant data from over 16 different types of documents in Salesforce to aid audits.

Description: To be able to provide non-emergency medical transportation catering to an array of vulnerable population sectors, SafeRide Health partners with a network of transportation vendors , but not before carefully vetting their backgrounds. They had a manual process where they took up to 16 different types of documents from each driver, including their vehicle registration and insurance documents, as well as their first aid/CPR training certificates, wheelchair securement training certificates, etc. Once processed, they manually entered relevant data points for each driver/vehicle based on their vendors in their Salesforce account.With the increase in demand, this process was proving to be cumbersome and not scalable. This prompted them to seek an OCR vendor that could automate this process reliably and in a cost-efficient manner, allowing them to scale their vendor network management process.

With our generative-AI powered OCR models, we were able to simplify and automate this process from end-to-end. It came down to the following 5 critical steps:

1. Import of documents: All vendors securely send their documents through ShareFile, which are automatically processed by Nanonets.
2. Classification of documents: Nanonets’ classification model intelligently identifies each of the 16 possible types of documents shared and directs it to the relevant OCR model for the data to be extracted.
3. Data extraction: Nanonets’ trained custom OCR models identify specified data points from different types of documents and extract them with high accuracy.
4. Validation of files and flagging discrepancies: The files then pass through validation checks put in place to weed out discrepancies, based on logics defined by SafeRide Health. In case of an error, the file is flagged and a member of the SafeRide team is notified and the files with no errors are approved automatically. 80% of the total files fall in the latter category.
5. Export to Salesforce: The files with errors are redirected to a dedicated “pending” folder in Salesforce, to be manually vetted by the SafeRide team, whereas the files with no errors are mapped to their respective Salesforce folder.