Intelligent Document Processing: Key Challenges

Although many-high performing organisations during last decade have been investing heavily in optimisation of business processes through delivering different transformation programmes such as e2e reengineering, lean & six sigma, business rightsizing, etc, there are still a lot of business processes which are paper-based driven and which have been for years were left out of automation programme scope. The list of paper-based driven processes requiring intelligent automation across different industries is quite huge, and still there are a lot of processes including from the list below are not yet fully automated across the industries:

  • Client / Supplier / Staff / Applicant onboarding
  • Client / Supplier / Staff / Applicant data verification
  • KYC (know-your-customer) verification
  • Compliance and Regulatory Client / Supplier data check
  • Invoice / Receipt Processing and Posting
  • Claim Processing

To greater understand the key challenges in scope of automating paper-based driven processes let’s consider the most common use case which could be applicable to any industry - “onboarding process” our company had a chance to automate for one of our client from finance industry through leveraging OCR, Computer Vision, Machine Learning and BPM & Document Management capabilities. We are listing below the key challenges and key observation points to be considered and which could be very helpful for any project team to examine before starting the project on intelligent document processing and automation.

Challenge No 1 (various type of documents)
Client or Supplier onboarding process is always associated with different type of the documents submitted by the Client or the Supplier: passport data, ID, different application forms, statutory declarations, constituent documents, etc. All documents submitted by the Client or the Supplier in scope of the mentioned process are of different type, form and layout by nature.
Consideration point: the target automated intelligent solution shall be capable to classify the documents by type and nature, capable to capture the required fields from the specific document by applying either template-based or templateless-based approach depending on image type and image complexity.

Challenge No 2 (bad quality of scanned documents)
Documents submitted by the Client or by the Supplier are of different quality and in most of the cases the quality of the scanned document is below the average. The documents submitted by the Client or the Supplier are being scanned either through mobile devices or by not high-quality scanners what generates the images with unnecessary artefacts, noises, wrong rotation.
Consideration point: the target automated intelligent solution shall be capable to clear the image from unnecessary artifacts, rotate the image, scale where required.

Challenge No 3 (multiple-pages documents)
Most of the documents requiring intelligent automation is multiple-pages documents where content or tables with content spans across different pages what complicates retrieving the correct data from the document.
Consideration point: the target automated intelligent solution shall be capable to understand that queued up for processing document of multi-page nature and required data could be located on any subsequent page.

Challenge No 4 (insufficient data accuracy recognition rate)
Most of the OCR engines which are being leveraged for recognizing the data from scanned documents are below acceptable threshold which makes difficult to achieve at least 80%-90% of automation.
Consideration point: along with selecting what OCR engine to be leveraged for your Intelligent Automated Solution, it is also required to supplement the OCR capabilities with additional advanced stack of technologies such as Computer Vision and Machine Learning to achieve greater results and higher level of automation.

Challenge No 5 (access rights management and security)
Not all business users within the organization shall have access rights to the documents scoped for automation. Access rights shall be granted granularly per type of the document and / or in accordance with roles & responsibilities set in the organization.
Consideration point: the target automated intelligent solution along with IDP (Intelligent Document Processing) capabilities shall also have access rights management capabilities which shall ensure the highest security and access control standards while deploying the intelligent automated solution within the organization.

There are many more other challenges awaiting ahead the intelligent automation project - leveraging right stack of technologies and technical capabilities could enable achieving the greatest success and highest level of automation for intelligent document processing.

Willing to know about more Intelligent Document Processing (IDP), please visit the following page: IDP - elDoc .

About «DMS Solutions»
«DMS Solutions» is a Technology company delivering Intelligent Automation Solutions. «DMS Solutions» is a vendor of Intelligent Document Processing & Document Management Solution (elDoc). «DMS Solutions» is your professional service partner in the field of Intelligent Automation and Advanced Robotic Process Automation. We leverage Machine Learning and Artificial Intelligence to build a powerful digital workforce for your business to win on the market. We focus on exploring new ways to apply disruptive technologies and recent inventions to bring innovative automation solutions in. We break boundaries and help clients to achieve their strategic goals through delivering next-gen solutions.
We operate globally covering US, EU, and APAC markets and we have offices in Hong Kong and Central East Europe - Ukraine.