About the Client
A tax resolution firm based in Jacksonville, Florida, receives tens of thousands of documents each month from the IRS and directly from clients. The firm needed a more efficient way to process those documents, link them to the correct CRM records, and route them into the appropriate servicing workflows.
Service Area
Document Management System (DMS) & Workflow Automation
Tool Stack
Custom DMS, OCR engine, Falcon Red mailroom machine, CRM integration
The Challenge
- Tens of thousands of documents arrived monthly from the IRS and directly from clients.
- Sorting, scanning, and linking each document to the right CRM account was manual, and manual errors were growing into a real business concern.
- A Falcon Red machine already opened, scanned, and tagged mail with basic metadata, but processing those documents was still slow and manual.
- In tax and accounting, even small delays can cost the firm and its clients dearly.
The Approach
Built a custom OCR-driven DMS to streamline and partly automate processing, designed around three document types:
- Over 30% of the documents received from the IRS department contained 3D barcodes, which contain four different data points about the document. The letters included the addressee’s personal information in the first half of the page. Using OCR, we extracted this information and created the appropriate records in the CRM. The file was moved under the DMS without any manual intervention.
- Approximately 40% of the documents had a 2D barcode with information about the type of letter with limited data points. The client’s identity information was also in the first half of the first page, which OCR extracted and linked to CRM and DMS. In the first version of the solution, a custom workflow was created for these documents so that an operator could fill in the missing info, which had to be extrapolated from the document content. Eventually, we automated much of this manual work for most documents, with only a small percentage needing manual intervention.
- One problem was that 30% of documents had no set structure and came without barcodes. These went through the OCR process to extract unique client identifiers and match them to CRM accounts. If a match was found with 100% probability, these were automatically categorized, else they went to the classification and linking workflow with suggested matches, and a mailroom operator would manually process the docs.
Based on classification and content, documents triggered servicing or sales workflows at varying priority levels in the CRM whenever action was needed.
Why OCR Was Important
The firm’s mailroom already had a machine that could open envelopes, scan documents, and apply basic metadata. The remaining challenge was making the scanned information useful for downstream processing.
OCR helped extract identifying information from document content so the system could attempt to match documents to the correct CRM records, classify them, and route them into the appropriate workflow.
Before vs. After
| Before | After |
|---|---|
| Manual sorting, scanning, and CRM linking | OCR-assisted processing and automated linking for eligible documents |
| Growing processing errors | Classification and matching workflows with human review where needed |
| Unstructured documents required more manual effort | OCR-based matching with suggested matches for uncertain documents |
| Document processing was slow and cumbersome | Faster processing and reduced manual handling |
| Delays in processing could affect client servicing | More timely document processing and servicing |
The Result
- 25% increase in upselling and cross-selling opportunities for existing clients.
- 80% reduction in document processing time following the DMS and OCR implementation.
- Reduced manual document handling and associated processing overhead.
- Faster document processing supported more timely client servicing.
Managing a high volume of tax documents?
Nablasol can help assess your document intake, classification, and CRM-linking workflows to identify where automation could reduce manual processing.
