Write every thing about Digital-TransformationI attached a Digital-Transformation-Report-2017.pdf prepare a 10-minute presentationYour will act as an instructor and “teach” your peers the concepts (The instructor will facilitate discussions)
digital_transformation_report_2017.pdf
Unformatted Attachment Preview
Intelligent Capture
For Digital Transformation
December 2016
Kemsley Design Ltd.
www.kemsleydesign.com
www.column2.com
Data capture from paper or electronic documents is
an essential step for most business processes, and
often is the initiator for customer-facing business
processes. Capture has traditionally required human
effort – data entry workers transcribing information
from paper documents, or copying and pasting text
from electronic documents – to expose information
for downstream processing. These manual capture
methods are inefficient and error-prone, but more
importantly, they hinder customer engagement and
self-service by placing an unnecessary barrier
between customers and the processes that serve
them.
Intelligent capture – including recognition, document
classification, data extraction and text analytics –
replaces manual capture with fully-automated
conversion of documents to business-ready data.
This streamlines the essential link between
customers and your business, enhancing the
customer journey and enabling digital transformation
of customer-facing processes.
1
The Digital Transformation
Imperative
Digital transformation – the fundamental
shift of business models and activities onto
digital platforms – requires rethinking and
retooling an organization’s end-to-end
business processes. Not just impacting
internal workers, digital transformation can
only occur through significant enablement
of customer channels. This requires
comprehensive self-service capabilities,
where customers create and complete
transactions without intervention from
workers inside the organization.
For example, a customer initiates a
property insurance claim by submitting a
notice of loss via their mobile device,
including photos of damage and supporting
documentation, and receives an autoadjudicated payment within minutes. Or a
customer applies for a personal loan by
completing an electronic application form
and uploading digital versions of their proof
of employment and pay advice, resulting in
same-day funds transfer.
Key to these scenarios is automation of the
internal processes and decisions, such as
auto-adjudication and payment processing;
plus the ability for the customer to directly
capture information into the processes,
including scanned paperwork, photographs,
filled PDF forms, and secure PDFs from
third parties. As the customer uploads
documents, smart recognition classifies
them, extracts data for interfacing with the
downstream processes, and identifies
missing information for additional uploads.
Straightforward transactions are completed
immediately without human intervention,
while more complex cases are routed to an
internal worker for resolution.
This is a fundamental shift in how
customers can choose to interact with your
organization through digitally transformed
channels, enabled by intelligent capture.
Are You Adding Roadblocks
To Your Customer Journey?
These scenarios might sound like a fardistant future, but they’re not: companies
are already doing this with today’s
technologies. Business process
management systems (BPMS) and related
technologies such as case management
(CM) and decision management (DM) are
commonly used for automating processes
within organizations. However, although
much attention is paid to modeling end-toend business processes and customer
journeys, often only internal activities are
automated.
What’s missing is the customers’ ability to
seamlessly interact with the processes,
including capturing information required to
2
drive them forward. Documents – whether
paper or electronic, structured forms or
free-form text – still serve as a primary
method for customers to provide
information, and this is typically used as a
cut-out point between the customer and
the internal process. Customers submit
their documents, but instead of intelligent
automated capture, the documents are
manually reviewed and processed by
internal workers to extract the required
information and add it to the internal
processes. Meanwhile, the customer is left
unhappily wondering what’s happening.
In the self-service scenarios described
above, the customer initiates the
transaction and sees it complete right
before their eyes, not days or weeks later.
However, if the actionable business data
can’t be captured from customer-submitted
information near-instantaneously, it
becomes a major roadblock in the
customer journey.
Winning At Efficiency And
Customer Satisfaction
Digital transformation is a great reason to
implement intelligent capture and process
automation, and may impact the success –
or even the survival – of your organization.
But it’s not the only reason: automated
intelligent capture has a number of direct
efficiency and customer satisfaction
benefits, even if it’s just for incremental
improvement of your business operations.
In addition to customer-facing processes, it
can also improve any points of integration
where documents – paper or electronic –
may change hands, including with business
partners and other internal departments.
In spite of the push to streamline business
processes and reduce paper, there are still
lots of documents in our processes, both
paper and electronic. As described in the
scenarios above, documents contain
essential data that drive business
processes and enable automation, such as
customer account identifiers and
transaction types.
Intelligent capture automatically extracts
this information from documents for use by
the downstream business processes. If the
documents are originally on paper, this
applies text recognition technologies to the
documents after scanning, including
recognition of machine-printed text (OCR),
hand-printed characters (ICR), handmarked forms (OMR), and barcodes.
Document classification determines the
type of document – an application form
versus a complaint letter, for example –to
ensure that the correct information is
captured and routed. [See the sidebar at the
end of this document, Fundamentals Of
Intelligent Capture, for more detail on these
technologies.]
3
Intelligent capture enables significant
efficiency impacts on business processes:
It reduces the data entry resources
required for documents provided by
customers and other external parties.
Even if paper is not involved, intelligent
capture from electronic documents can
avoid manual transcription or copy-andpaste activities.
It reduces the cycle time of processes
by eliminating the time required for data
entry.
It increases data accuracy over manual
data entry, reducing rework and
improving process quality.
Adding customer self-service to this can
improve internal efficiencies even more,
since customers do their own data entry
and scan their own documents.
This can radically change staffing
requirements, with internal resources now
free to perform value-added activities to
handle complex exception cases and
service customers. Data entry outsourcing
or offshoring may be completely
eliminated, although scanning operations
would still be required for customersubmitted paper documents.
capture has performance benefits
throughout the entire business process. By
injecting captured data directly into
business processes, it enables automated
decisioning and straight-through
transaction processing, further reducing
process cycle time and manual efforts.
With intelligent capture and smarter
downstream processing, entire processes
that were previously outsourced become
targets for re-insourcing.
The result: customers are happier with
faster, more accurate processes and selfservice options, driving the potential for
increased revenues. Business executives
are happier with more efficient and higher
quality business operations, which reduces
costs and can improve compliance.
Summary
With paper and electronic documents still
widely used for customer interactions, it’s
clear that smarter data capture is essential
to smarter business processes. Intelligent
capture can reduce costs and cycle times,
but more importantly, can enable digital
transformation of customer-facing
processes.
Making the front end of processes more
efficient and accurate through intelligent
4
Sidebar: Fundamentals Of Intelligent Capture
Intelligent capture encompasses a collection of technologies that can be used together or independently,
depending on the source of information. Documents – broadly defined as “any recorded information or
object which can be treated as a unit”1 – may have originated on paper or created in a machine-readable
electronic form; their content may contain structural elements such as fields, or may be completely
unstructured.
Converting Paper To Digital
A paper document is first converted into a digital image using a scanner or, more commonly when a
customer captures the document, a smartphone application that uses the device’s camera. The digital
image is simply a picture of the paper document, with every crease and smudge faithfully reproduced, but
no knowledge of what information the document contains. As part of the scanning process or in postprocessing, a variety of recognition techniques may be applied to extract the content into editable text:
Optical mark recognition (OMR), which detects human-marked fields on a form such as a test or
survey.
Barcode recognition, which translates barcodes into data.
Optical character recognition (OCR), which recognizes machine-printed characters.
Intelligent character recognition (ICR), which recognizes hand-printed text.
OMR and barcode recognition have effectively 100% accuracy. Barcodes in particular are ideal as
document identifiers: a customer application form that is filled out online then printed for signature can
dynamically encode the customer ID and other data in a barcode on the document. Once scanned, this data
is used to classify the document and pair it with an internal process.
OCR has near-100% accuracy under perfect conditions: if the document is not physically damaged, the
scan is good quality and the font size is sufficiently large, machine-printed characters can be recognized
reliably. ICR has much lower accuracy for hand-printed characters, although it can work well for restricted
character sets such as postal codes or numbers, or within constrained fields so that the characters don’t
touch.
Document Classification
Once a paper document is scanned and converted to editable text, it is very similar to a natively-digital
document such as a word processing document or a filled PDF form. At that stage, regardless of origin,
document classification is used to identify the type of document based on general structure or other
identifiers. For example, a government form can be identified by the form number or barcode found in a
fixed location, while a free-form letter requires more complex structural analysis to identify components
such as an address block, text paragraphs, and a closing block. Natively-digital documents may also
include classification information in the metadata.
Classification is an essential part of intelligent capture, since the type of document will determine what
downstream processing is applied. It is also required to separate unique content from the background
(preprinted) text in form documents. A variety of techniques are available for classification, from simple
pattern matching to sophisticated natural language processing.
1
ISO 12651-2:2014, “Electronic document management – Vocabulary”, International Standards Organization.
5
Text Analysis And Extraction
Once a document is fully machine-readable, and the document type has been identified, information from
the document is extracted in the context of the business process.
For a structured form where the form type is recognized, this is usually straightforward: the form
background is ignored and the field data can be easily extracted and matched with data fields in related
processes. For example, a US government W-9 tax form always has the taxpayer’s name in the first field of
the first section, and the social security number (SSN) in the first field of the second section, regardless of
whether the form was filled out by hand, with a typewriter, or as a fillable PDF form. In fact, fields can be
used to cross-check each other if the recognition accuracy is in doubt: a handwritten account number is
more easily recognizable due to the limited character set, then can be looked up in a line of business
system and used as a cross-reference on the handwritten name field.
Structured forms of unrecognized types are more difficult, but may have enough features to allow
sophisticated text analytics engines to recognize the intent and content of the document. Although many
systems require manual training on different possible document layouts, some can learn new layouts
automatically. A common business-to-business example is invoices, where most accounts payable
processes still require manual data entry from a vendor’s unique invoice format into the internal payables
system. In the consumer-facing scenarios described earlier, pay advice forms from different employers can
assume many different forms, but are all proof of employment so will have commonly recognized fields
such as salary amount.
Unstructured documents present an even greater challenge, since the information of value may be
sprinkled throughout the document. A customer may write an email that describes a purchase that they
want to make and provided all required information, but it can be difficult to recognize the content within
the correct context: is a number in the text the size of the object or the number of objects to be ordered?
Photo Classification
Photographs are a type of unstructured document, but use very different classification and data extraction
techniques than for text-based documents. Some of the more successful techniques for photo
classification involve deep learning, which attempts to model high-level abstractions for objects in the
photographs. This may be supplemented by crowdsourcing, where labels applied to publicly-available
photographs are used to create a large training set.
6
About the Author
Sandy Kemsley is an independent analyst, consultant and process architect specializing in
business process management and the social enterprise. During her career, she has
founded companies in the area of content management, process management and ecommerce, and held the position of BPM evangelist for a major software vendor.
Sandy writes a popular BPM blog at www.column2.com and is a featured conference
speaker on BPM and digital transformation. She is a contributing author to books on
social BPM and adaptive case management, and the winner of the 2016 Marvin L.
Manheim award for significant contributions in the field of workflow.
Sponsor
This white paper was sponsored by ABBYY.
ABBYY creates technologies and solutions that enable people to action information. As a
worldwide company that sets the standard for content capture and language-based
technologies, ABBYY software integrates across the entire information lifecycle to
transform data into useful information – enabling organizations to optimize business
processes, mitigate risk, accelerate decision-making and drive revenue. With 14 global
offices and a team comprising over 1,200 people, including over 300 industry-leading
engineers, ABBYY’s innovative culture has earned it over 250 major awards and
partnerships with major manufacturers of document capture and mobile devices. Its
technology now serves over 30 million individuals the world over – enriching people’s lives
and empowering businesses of virtually every size. From home offices to multinational
enterprises, ABBYY streamlines workflows by automating time- and labor-intensive tasks.
www.ABBYY.com
7
…
Purchase answer to see full
attachment
