Introduction
Each and every capture process involving
paper has the same problem. How to reduce the cost of scanning and data entry?
With Square 9 Advanced Capture users
now have the ability to compile many types of paper
documents without separater sheets, place them into
the scanner and have a technology classify but also
intelligently extract non-templated OCR data. Moreover,
while it classifies and extracts data, it also integrates to the SmartSearch document
management system for storage at the same time as
it feeds another
line of business application
for adjudication purposes.
Reference 1:
Classification is performed by keyword search from
the OCR output.
Spatial relationships between keywords as well as
hierarchical classification allows thousands of types
of documents to be captured
and interpreted.
Reference 2: The classification
process behaves like a mail room person - rough sort first, then a finer sort. Leveraging “Following
Page” detection, removes the need for separator sheets
between documents and rules for detecting “following pages”
are configurable.
What Is It?
Square9 Advanced Capture is a platform
for the development of applications for document classification and data extraction leveraging best-of-breed technology
solutions from Paradatec's Prosar-
Aida, Captaris' DokuStar, or
Kofax Transformations.
■ Advanced Capture is an unstructured data capture product that can
intelligently extract and classify
documents using revolutionary
artificial intelligence (AI)
capabilities. This solution will allow between 70%-96% capture automation, meaning the percentage of documents that flow through Advanced Capture without the need to be manually corrected.
■ Square9 Advanced Capture is the solution for the common capture problems that companies face when implementing document management–which is how to control the cost of capture and data entry.
■ It can be used to tackle all kinds of input management tasks, regardless of whether the volume of documents is high or low, the workflow is straightforward or complex, the processing is central or distributed, or if the documents to be processed are structured or unstructured.
Classification and Data Extraction
■Advanced Capture is text recognition and
analysis software based on artificial intelligence
that can
ascertain document types and
classify images for appropriate
routing and archival.
■ The process begins with a highspeed
and efficient full page OCR
scan of each image. This first step
allows Advanced Capture to search
each and every word of every
document to discover the document
content.
■ In order to classify the current
document, Square9 Advanced
Capture searches the document for
tell-tale characteristics specified by
the administrator using pattern
matching, spatial relationships and
synonyms like invoice number,
invoice #, inv_num, etc. These tend
to be certain keywords or phrases
■ This overall process of discovery is
unique and makes Advanced
Capture very different from the
typical template based solutions.
With other solutions the document
content is expected and assumed to
always appear in the same
geographical areas.
■A significant advantage of Advanced
Capture document content
discovery method is its ability to
process virtually unlimited number
of versions and formats of particular
document types.
■Advanced Capture is confidence
based, meaning if the Advanced
Capture engine does not have a high
degree of confidence, the data value
is flag and sent for correction within
the data validation module.
■ Since the Advanced Capture engine
is transparent to the user, the only
interaction the user has with the
system is within the validation
module, simply correcting values
and rerouting them back into the
process workflow.
Reference 3:
During the extraction process label-value pair logic
is used to find
data. Utilizing the extracted OCR, the solution
finds all synonyms of
a type like “invoice date”, “inv date”, “date”. Then
the artificial intelligence engine discovers the
date
that is spatially in the
correct RELATIVE location to the label. Note – no
fixed zones are used – anywhere!
Reference 4: Advanced Capture supports tabular data
extraction for any type
of document: EOB, invoice, traffic instructions or
sales orders. Each
type of document is processed same way, with no limit
to the variety of layouts
Application Use Scenario's
Square 9 Advanced Capture can be used in
many business applications but typically its used in a few specific verticals, namely: AP, healthcare, mortgage, media broadcasting and manufacturing.
■ Automate AP invoice, EOB and sales order entry – capture sales orders from fax, mail, email or print and automate the process of classification, data entry as well as
integration with internal line-ofbusiness systems
and document management archiving.
■ Automate traffic instructions – media
outlets like cable, radio and satellite receive large amounts of traffic instructions which tell the media provider which commercial
to run, how long, on what station and at what frequency.
■ Document classification and distribution
in mail system.
■ Extraction of business process relevant data
from documents.
■ Capturing unstructured forms, surveys, assessments
and exams. 
Use Case
• By improving the speed and efficiency at which a
company captures and processes documents such as patient
Explanation of Benefits (EOB) information, Square9
Advanced Capture can produce measurable benefits at
every level of the organization. The benefits resulting
from unstructured data capture encompass additional
revenues as well as savings from reduced costs. The
savings are real- hard cost dollars from reducing
an enormous data entry staff both on the payables
and document management application side.
Hard-dollar benefits include:
Labor savings
• Reduced need for data entry operators; document
prep workers, sorters, etc.,
• Increased accuracy of data extracted from EOB’s,
invoices, loan documents reduces the time spent on
error correction
Improved cash management
• Faster payment postings and access to more accurate
payment data, means increased control over billing
and accounting processes, which leads to better overall
cash flow management
• Faster, more accurate generation of critical payment
data, expedites the process of Secondary Claim submissions
Rapid return on investment
• In mid to high volume organizations, the payback
on unstructured data capture is less than a year.
Business Case
• One of the largest healthcare providers on the
west coast, spread across nine states, incurred
an enormous
cost processing and indexing EOB’s into the existing
patient billing adjudication and document
management software's.
• With the shear number of EOB’s the hospital processed
daily, the 20 FTE were being augmented with 15 full-time
temps.
• The bottleneck in the process was separately indexing
the EOB into the host document management system
and into patient billing software. Additionally
since the layout of the EOB was wide ranging and
difficult
to read,
there
was no good solution to accurately indexing the
EOB
into the document management solution, so the healthcare
provider settled on an inefficient full text index
schema. As a result thousands of EOB’s are lost each
day and productivity of FTE’s was further impacted
by the full text engine that required 1.5-2.0 minutes
processing time per EOB.
• With 20 FTE’s and 15 FT temps, the healthcare
provider still had a significant backlog of EOB’s
which required overtime night and weekend processing
which escalated staff costs. Based on a strategic
assessment, it was estimated that they needed 50 FTE’s
at a cost of $160,000 per month to stabilize the processing
of the current EOB backlog and this estimate did not
include future growth.
• The patient financial services (PFS) department
decided to implement an unstructured data capture
solution to fully meet the patient billing backlog
• The ROI on this solution was 1.5 months and only
required a PFS FTE count of 6 down from 20 and no
requirement for additional staff augmentation. |
