The Indexing Awakening
SPUD gave its documents a home. But Orson the Onion is still opening every loan packet, separating every file, and indexing every page by hand.
The documents have a home. The manual work followed them there.
Cycle One gave SPUD Credit Union a central place for its content. Employees no longer have to search disconnected folders and personal inboxes to find a loan document.
But centralizing the files did not change what happens when a new loan packet arrives. One PDF may contain a loan application, bank statements, utility bills, credit reports, identification, and other supporting documents.
Someone still has to determine where each document begins and ends. Someone still has to identify every document type. Someone still has to read the pages and type loan numbers, member numbers, dates, income, and other information into the system.
That someone is Orson.
Centralizing content made the documents easier to find. It did not make them easier to process.
Teach the system to understand what arrived.
To remove the repeatable work from Orson's queue, Cycle Two has five assignments.
- 01
Determine what each document is.
Identify loan applications, bank statements, credit reports, utility bills, and other expected document types.
- 02
Separate the document package.
Turn one combined PDF into the individual documents contained inside it.
- 03
Capture the data.
Extract loan numbers, member numbers, dates, income, account information, and other configured values.
- 04
Understand the context.
Gather useful information that may not appear inside a neatly labeled field.
- 05
Put the information to work.
Send the separated documents and captured information into the content repository, ready for the next process.
From manual indexing to intelligent document processing.
Before: manual indexing
- One combined loan packet arrives as a single PDF
- An employee opens and reviews every page
- The employee decides where each document begins and ends
- Every document type is assigned manually
- Loan and member information is typed into fields
- Context must be assembled by reading the complete package
- Routine processing consumes time that could be spent on exceptions
After: intelligent document processing
- One combined loan packet enters the process
- Individual documents are separated automatically
- Each document is classified by type
- Configured business information is extracted
- Contextual summaries and suggestions support the review
- Documents and captured information are added to the repository
- Employees focus on exceptions and decisions that require human judgment
Cycle Two does not remove people from the process. It removes the need for people to manually process every page before useful work can begin.
One packet goes in. Organized information comes out.
Hyland Intelligent Document Processing uses the document types and keyword information already configured for SPUD's content repository. That existing structure helps the system recognize the documents SPUD expects to receive.
When a combined loan packet arrives, the system examines its contents, separates the pages into individual documents, and classifies each document by type.
It then extracts configured values such as loan number, member number, branch, dates, and other business information. Reasoning fields can also create summaries or provide suggestions based on the document's context.
The resulting documents and information are sent into the repository, ready to support the next part of the loan process.
- Step 1
Recognize
Determine which document types are present inside the incoming packet.
- Step 2
Separate and understand
Split the packet, classify each document, extract configured values, and gather useful context.
- Step 3
Put it to work
Send the documents and captured information into the repository for review and downstream processing.
Automation does not eliminate exceptions or replace lending decisions. It gives employees more time to review exceptions and apply SPUD's policies.
Watch SPUD automate the work between arrival and filing.
This walkthrough shows how one combined loan packet can be separated, classified, enriched with useful information, and added to the content repository.
Cycle Two Demo: Separate, Classify, Extract, and Understand
Approximately 11 minutes and 30 seconds
See how SPUD uses intelligent document processing to identify documents, split a combined loan packet, extract configured information, create contextual suggestions, and send the results into its content repository.
What you will see
- 01
- 02
- 03
- 04
- 05
- 06
- 07
- 08
- 09
- 10
Orson can stop indexing every page.
The next time a combined loan package arrives, Orson does not have to manually split the PDF into individual documents. He does not have to identify every page or type every configured value into the system.
The repeatable work happens first. The loan application, bank statements, utility bills, identification, and supporting documents emerge as separate, classified records with useful information attached.
Orson still matters. He reviews exceptions, checks the information that needs attention, and applies the judgment that belongs with a person.
SPUD's content is no longer simply centralized. It arrives organized, classified, and understood.
- Documents separated and classified
- Business information captured
- People focused on exceptions
What happens after the documents are understood?
SPUD can now receive organized content and usable information. In Cycle Three, that information begins moving through the work itself.