A global visual technology company was paying for costly third-party software to scan thousands of documents, and the work was slow and resource heavy. Here is how our OCR software development team approached it.
Client
EMS Imaging, a global visual technology company
Industry
Visual technology and document imaging
Services
Custom OCR software development, document digitization, scanning, archiving and data warehousing, web portal access
Scope
A rules-based OCR engine with algorithmic cataloging and, more recently, deep learning with AI
Headline result
14% ROI in the first year (as listed on the Digital Success case studies page)
Digital Success built an OCR solution for EMS Imaging to replace costly third-party scanning software. The system digitizes, archives, catalogs and organizes records using a rules-based engine, algorithms and deep learning with AI. It reads handwritten documents in different languages, protects confidential information, and makes records searchable through a web portal.
14%
ROI in the first year, as listed by Digital Success
3
core capabilities: OCR scanning, secure portal access and metadata search
The challenge: costly software and a tedious scanning process
EMS Imaging used costly third-party software to scan thousands of documents. The scanning effort was tedious and resource intensive. The client needed an OCR solution that could digitize, archive, catalog and organize its records.
Our approach: custom OCR software development in three parts
We built the system around three documented capabilities. The table summarizes what each one does.
| Capability | What we did |
|---|---|
| 1. OCR scanning and archiving | Built an OCR with a rules-based engine for scanning, archiving and data warehousing. The engine digitizes and catalogs documents accurately using algorithms and, more recently, deep learning with AI. It can scan handwritten documents in different languages and have the system understand them, at high speed, ease and accuracy. |
| 2. Secure access | Protected confidential information securely and made it accessible to customers through a portal available over the web. |
| 3. Search and metadata | Improved search and produced accurate metadata so that documents can be found quickly and the data is easy to access. |
A design choice we made: start with rules, then add AI
The solution began with a rules-based engine and added deep learning with AI later. Rules give predictable, checkable results for document types that follow a known layout. Deep learning adds flexibility for material that does not, such as handwriting and mixed languages. Layering the two, rather than replacing one with the other, is how the system covers both kinds of document.
How to read the sections that follow. Each section carries a tag. Campaign evidence means activities and results documented in this engagement. Industry context and recommended practice are general guidance from Digital Success; they were not measured in this campaign and are not results.
Why Document Digitization Is More Than Scanning
CAMPAIGN EVIDENCE + INDUSTRY CONTEXT
A scanner turns paper into images. It does not make those images findable, sortable or safe to share. The client’s stated need went further than scanning: it wanted records digitized, archived, cataloged and organized. That is a data problem as much as an imaging one.
This is why the documented solution includes data warehousing, metadata and a web portal alongside OCR. A record that is scanned but not cataloged is still hard to retrieve. The case study’s focus on search and accurate metadata reflects that.
Still paying for scanning software that does not fit?
Tell us about your documents and volumes, and we will outline what a custom OCR and archiving solution could look like.
Document Scanning and Archiving Services: Building a Searchable Record System
CAMPAIGN EVIDENCE + RECOMMENDED PRACTICE
Documented: the team built an OCR with a rules-based engine for scanning, archiving and data warehousing, added deep learning with AI, protected confidential information, provided a customer web portal, and improved search with accurate metadata.
Recommended practice for any scanning and archiving project:
- Decide the metadata first. Agree which fields every record needs, such as type, date and customer, before scanning starts.
- Test on your worst documents. Include handwritten, faded and multilingual samples in the pilot, not only clean printouts.
- Measure accuracy on a sample. Check a fixed sample of records by hand and track the error rate over time.
- Plan access and security together. Decide who can see which records before the portal goes live.
Need your records searchable and secure?
Share the document types and volumes you handle and we will plan the scanning, archiving and portal approach.
What Other Teams Can Learn from This Project
CAMPAIGN EVIDENCE, THEN RECOMMENDED APPLICATION
| Lesson | What happened here (documented) | How to apply it (recommended) |
|---|---|---|
| Start from the real cost | Costly third-party software and tedious scanning drove the project | Calculate current scanning cost before choosing a solution |
| Digitize and organize | Cataloging, metadata and data warehousing were part of the build | Define how records will be searched before scanning them |
| Layer rules and AI | A rules-based engine came first, with deep learning added | Use rules for predictable documents and AI for the hard cases |
| Cover hard documents | Handwritten and multilingual documents were supported | Include them in testing from the beginning |
| Protect and share | Confidential data was secured and offered through a portal | Set access rules before opening the portal to customers |
Key takeaways for OCR and document digitization
- Scanning is the start, not the finish. The solution covered cataloging, data warehousing, metadata and search, not only OCR.
- Rules and AI work together. A rules-based engine was extended with deep learning.
- Security and access are part of the design. Confidential records were protected and made available through a customer portal.
Frequently Asked Questions
Digital Success built an OCR solution with a rules-based engine for scanning, archiving and data warehousing. It digitizes and catalogs documents using algorithms and, more recently, deep learning with AI, and gives customers web portal access to their records.
EMS Imaging was using costly third-party software to scan thousands of documents. The scanning effort was tedious and resource intensive, so the client needed a solution to digitize, archive, catalog and organize its records.
According to the case study, the OCR can scan handwritten documents in different languages and have the system understand them, at high speed, ease and accuracy.
Digital Success lists a 14% ROI in the first year for EMS Imaging. The cost basis, volumes, accuracy rates and time saved are not published in the text of the case study.
Need a custom OCR or document archiving solution?
Our team will review your documents and workflow and show you what to build first.
Talk to our team Call +1-855-344-4821