Effective and efficient distributed, web based document capture architecture

Effective and Efficient Distributed, Web‐Based
Document Capture Architecture
With Tricks, Tips and Realities of the Capture Process

By Paul E. Szemplinski

A White Paper from CAPSYS Technologies, LLC

© Copyright 2009 CAPSYS Technologies, LLC. All Rights Reserved.
Version 1.0 1 www.CAPSYStech.com

Introduction

In contrast to Centralized Document Capture that requires all documents to be scanned with
hardware and software at one central location, Distributed or decentralized Document Capture
software (or Software as a Service) is a welcome, cost‐effective alternative.

To date, Document Capture offerings in a Distributed model have been extremely limited.
The distributed model impacts the traditional Centralized Capture licensing revenue
generated by many major software makers.  Those per‐click counts, seat and server licenses
add up!

Distributed Capture systems allow companies to shift the usual capture costs outward to
the point of document origin – eliminating shipping, software, maintenance, labor and
other costs.  With newly‐available Distributed Capture over the Web, companies can shift
costs almost entirely outside of the organization, bringing even greater savings.

Most importantly, these savings can be achieved with Distributed Capture software/SaaS
systems without sacrificing the capabilities of traditional software.

This white paper describes the fundamental building blocks of Distributed, Web‐Based
Document Capture architecture, and share numerous tricks, tips and realities of the Capture
process:

• Acquiring documents………………………………………………..Page 4
• Associating document pages…………………………………….Page 13
• Assign Metadata……………………………………………………….Page 18
• Automate as much as possible………………………………….Page 22
• Address exceptions…………………………………………………..Page 27
• Archive to the Repository ………………………………………..Page 30
• Analytics of the Capture Process ……………………………..Page 31
• CAPSYS CAPTURE Distributed, Web‐Based
       Document Capture……………..………………………………….. Page 33


Acquiring Documents


At first glance, the concept of document scanning seems trivial. I have a piece of
paper, it needs to be imaged; therefore I scan it, give it a file name and viola! My job
is done.

Eh, not quite that simple. Try taking such a simplistic approach when you are dealing
with tens, hundreds, thousands or even millions of documents. Large numbers of
documents require sophisticated software and powerful scanners just to keep up!
You can’t simply scan the document using the freebie stuff you get with your
favorite scanner manufacturer, treat images like you do your file sharing
applications and then dump it into your SharePoint system. It just won’t work.

Take a haphazard approach toward your Document Capture process, and you’ll end
up with a shoddy repository, regardless of the vendor’s claim of how wonderful the
features and functionality of ECM repository or search technology might be!

Remember, It’s Not Just about the Repository
Should all eyes, ears and the organization’s attention be solely on the ECM
Repository? We’ll argue not (of course) though the reason may not be quite so
obvious.

The content arriving at your organization from external customers and/or trading
partners has intrinsic value to the organization; otherwise you wouldn’t be
contemplating ECM in the first place. We’ll bet that, as that content arrives,
somebody in your firm takes action and uses it during the normal course of business.

For instance, when an invoice from your paper supplier
Without Document
arrives, the mailroom opens it, date stamps it and then
Capture, how do you
routes it to Accounts Payable, where it is then matched
intend to transform all
to the original PO or shipper code. A voucher is created,
and eventually the invoice moves back to the originating of that paper into an
requestor for approval before being routed back to AP for ECM repository?
eventual payment.

So in this example, is the mere fact a company might be contemplating imaging this
invoice for archiving purposes into a repository (i.e. SharePoint, Oracle or whatever)
of great value to the company, its shareholders, its customers or its trading
partners?

Frankly not, who really gives a hoot? And who in their right mind can justify the
expense of an electronic filing cabinet in the 21st century? Last time we checked, the


cost of a filing cabinet is still in the sub‐$100 range, and the cost of an ECM solution
is still, well, let’s just say slightly more expensive than $100!

To set the record straight, we’ll argue that as soon as that invoice is generated by
the supplier, it should be directed to your firm electronically if at all possible. (We
are not speaking of a facsimile representation, rather “electronic” meaning the raw
data.)

Many financial systems (such as Microsoft Dynamics GP, Oracle Financials, MAS 90,
etc.) can accept electronically submitted invoices and then process them for
payment. At the end of the process, you can “archive” a copy of the completed
transaction into your ECM repository. That, by far, is the least expensive route to go
because no scanning is required!

However, this level of process optimization is not always possible for a number of
reasons.

• What if you can’t get the information electronically because you don’t have
control over the delivery, or you require documents with handwritten
signatures or other forms of annotation as the final proof or evidence of
doing business together?
• What if your organization is geographically dispersed? Should you be
shipping the documents back and forth or faxing them to the central
location?

In any of these cases, you are presented with a challenge that neither the ECM
repository nor the ERP system in and of themselves can solve. Thus, you need a
Document Capture solution to augment your overall ECM strategy.


Hardware Choices: Multi‐Function Printer, Document
Scanner, Network Appliance Scanner or Facsimile?
Pick your favorite — fax, MFP, or scan — there are pros and cons to each. Here are a
few thoughts to ponder before you make your decision.

The MFP (Multi‐Function Printer)
You are all too familiar with it. You bought that honking piece
of equipment from one of the big guys, and there are a few of
them (or perhaps dozens of them) scattered around the
country at your various facilities. They sold you on all the bells
and whistles: it copies, it prints, it faxes, it even scans.

However, you probably haven’t quite figured out how that
scanning function works yet. There’s a nice executive‐looking
button on the panel, but where in the heck does that imaged document actually go?
How do I find what I scanned later to retrieve it? Does it get indexed? Can I retrieve
that document from my own workstation or the Internet once it is scanned? Ah, but
you are clever. You will just e‐mail that document to yourself and stuff it in your own
personal e‐mail archives. Now you have your IT guy screaming at you because your
e‐mail box has exceeded its quota!

The fact of the matter is the MFP has potential, but has it readily been adopted as an
acceptable alternative to the traditional document scanner? Not yet and here’s why:

• First, you have a department of people that must share that single device to
do copying, printing and faxing. Are they going to wait in line to use it for
mission‐critical document scanning applications like AP, claims, application
processing or new business underwriting?
• Are you happy with the feeder on your current copy machine?
• Are the speeds and quality of the document scans from a MFP comparable to
that of a workgroup or departmental scanner from Fujitsu or Eastman
Kodak? (No!)
• What about the fact that document scanners have built‐in image
enhancement for every scan. Images get nearly perfected before they are
presented for review. (Features like de‐speckle, de‐skew, auto color
detection, autorotation, cropping, punch‐hole removal, etc. are inherent in
most of today’s scanners and Capture software.) Good luck trying to figure
out all that advanced stuff on your MFP. (Better yet, call your local MFP
supplier and ask them to give you a tutorial on how to make all that stuff
work in a predictable manner each and every time you scan.)


No doubt, the quality of the MFP has been improving, and will continue to improve.
But, it took a decade or so for the scanner hardware and software manufacturers to
collaborate and earnestly put forth an engineering effort to adequately address
these challenging technical problems. It is still not perfect, but they are light years
ahead of the MFP guys.

What’s the moral of the story with MFPs? They are adequate for ad‐hoc document
scanning. If you need to do serious document scanning, meaning the majority of
your workload is document‐scanning related, you should buy the right tool for the
job: a document scanner. If MFPs could do everything a document scanner could,
would there be such a huge market for distributed document scanners?

Be prepared to have custom interfaces written or modules purchased (for a price) to
take the content imaged by the copy machine and import it into your Capture
Software platform. If you have MFPs and you have a limited budget, figure out how
to leverage them in your overall Capture process. In the short run, we predict that
you and your employees will be less than pleased with the MFP’s overall value
proposition, and you’ll eventually reach out into the market for a real document
scanner. (Yes, we know the hate mail will be flooding our mailboxes from the MFP
providers.)

Document Scanners
What a plethora of options to choose from — especially for the Distributed Capture
market!

The fact is, distributed Graphic placeholder
scanning is where the Which one of
majority of R&D following scanners
expenditures are going is up for the task?
from the major scanner
manufacturers. Why? This
is where you, the
customer, desire your
price point and functional
requirements to be!

You can get great quality
at an affordable price —
feature packed — with
distributed document
scanners from leading
manufacturers, such as Fujitsu, Eastman Kodak and Canon. Some vendors offer a
wide family of products that meet the needs of the desktop, workgroup, department
and the mainframe; others focus primarily on the desktop, workgroup or the


department level. Visioneer makes an awesome portable TWAIN, USB‐powered
color scanner (about the size of a rolling pin) for the true road warrior, and what do
you know, they appropriately named the product, “RoadWarrior.” And, it’s priced
around $200. All of these distributed Capture scanning products are designed to be
efficient and cost effective and are optimized for the demands of distributed
document scanning.

In our opinion, of the numerous choices you have in today’s market from a device
standpoint, two distributed scanning options are available to choose from:
1. Desktop/workgroup document scanners
2. Network scanning appliances

The two major contenders in both of the market segments (from a product family
standpoint) include Eastman Kodak and Fujitsu.

Figure 1
If you want your scanner attached
to your PC, Desktop/workgroup
scanners are where you should be
looking. The scanners range in
sizes, shapes, daily‐duty cycles and,
of course, the price. The interfaces
required to connect the scanner to
the PC are usually one of the
following: USB, Firewire or SCSI
(although declining in popularity).
USB certainly is the most popular
from a desktop, Distributed
Capture perspective. TWAIN

drivers usually ship with the
Distributed, Web-Based Document Scanning
scanners and are usually
interface.
developed by the scanner
manufacturer itself; thus the features of the scanners are likely to be fully exposed
via the TWAIN interface. ISIS is another alternative, but it usually comes with an
additional license cost and perhaps even royalty costs. TWAIN is generally free and is
bundled with the scanner at the time of purchase.

Network Scanners: Yeah Baby, Slam Dunk! Imagine this: In just moments, you could
share documents across your office network. A color LCD touch screen makes it easy
for anyone in your office to send documents to multiple network destinations,
simultaneously. (See Figure 1.) You can scan to e‐mail, to folders, to print, to
portable USB drives, to searchable PDF — all with one product that requires no PC
workstation or extra software! And your images will have the outstanding quality
that comes from an all‐in‐one appliance. Is it fact or fiction? It’s fact.


These network appliances are what our industry has been waiting for! When the
MFP was first introduced, it promised much, but failed to deliver. Finally, they are
here — the network scanner appliances. Everything you need is bundled and self‐
contained within the unit: a high quality, high speed, duplex document scanner
(black and white, color, gray scale, PDF, auto feeder, etc.), a PC, the scanning
application, a network cable, remote management capability, and it is really simple
to use! You can’t beat it from a price performance standpoint.

Network scanning Distributed scanning software companies have the ability
appliances should be to interface directly to the device via a software
seriously considered development toolkit (SDK), thus embedding their
as an integral
scanning application right on to the device. So when you
pull it out of the box, you plug it into the wall, and you
component of your
are ready to go with very minimal setup time (maybe an
ECM rollout strategy.
hour or two as you still have to deal with your internal
network issues).

Do your own cost comparison of building out a standard PC to support a document
scanning application (PC, Windows, scanner, monitor, internal support and
maintenance), and then compare all those individual items against the single cost of
an all‐inclusive network scanner appliance.

You can even remotely manage them through a central management application
back at the corporate office. There really is no comparison between the two options.
The network scanner appliances, by far, offer the very best ROI and lowest total
cost of ownership. Consumable costs and ongoing maintenance services are both
inexpensive, ease of use has been achieved — far superior than the MFP — and
perhaps even slightly better (in some instances) than a PC/scanner combination.


Fi-6010N

Network Scanners

Both the Fujitsu and Kodak color
network scanners are easy‐to‐use,
reliable alternatives to a

traditional scanner (as well as a

Multi‐ function Printer/Copier).  In
fact, we bet you’ll never use a
copier for scanning again!

Just load your documents into the
feeder, and use the large, color
LCD touch screen to send it to
your Capture application (i.e.

CAPSYS CAPTURE™ software),
Scan Station 500 Series network fax, printer or electronic

folder. You can even preview
images before sending them to e‐
mail.

The bottom line — if you are
contemplating implementing a
Distributed Capture solution,
these network scanners are an
option you should seriously
consider: They offer a small
footprint and a low price point
and produce much higher quality
scans than multifunction
copiers/printers.


Faxing
Tried and tested and they work just fine. Just
eliminate the paper faxes. If you are still using
paper fax machines, why? If you receive faxes in
paper format, you are then forcing yourself to deal
with the process of prepping, scanning and
indexing manually. That means you now have to
face the reality that you will need to attach a labor
force to the fax process, thus increasing your
overall cost of doing business.

Really, that fax machine is a relic and belongs in the Please, just say NO to
Smithsonian or recycling bin. It is probably costing paper faxes!
you more money to operate than you think. Do
yourself a favor and simply turn it off and introduce
your firm to the concept of an electronic fax server
or even a subscription‐based fax service.

Graphic placeholder
Maybe you did not
know that you can
automatically capture
your faxes electronically
with little or no labor
involved and do it all
reliably and
inexpensively? Plus,
that same automated
fax platform that is used
to accept incoming
faxes electronically can
be used to send out
faxes electronically
without having to first
print the document off,
manually create a fax
cover sheet, get up
from your desk to walk
over to the fax machine,
dial a number, and send
and wait for the
confirmation.


With a Distributed, Web‐Based Document Capture application, you can add in a pre‐
built module that will automatically “poll” at predefined intervals the incoming
electronic fax queue, and viola! Your faxes have been auto‐magically (technical
humor) transformed into an image that can be further processed inside your
Capture Process flow.


Associating Document Pages

Here is a something to ponder: If you went down the road of centralized scanning,
thus accepting the fact that you are automatically increasing your operational costs
by agreeing to such a decision, think now about the following choices you have to
make:

• You have all that paper coming into your facility. You need plenty of staff and
work space to prep the documents (remove staples and paper clips, classify
or sort documents/pages into logical groupings, combine them into bundles
of a 100 to 500 pages per batch, etc.)

• Don’t forget that you have to come up with a means to tell the software
application how to distinguish the beginning of one document from the
beginning of the next. Normally, this means you have to introduce the
concept of document separators (usually blank pages, bar codes or patch
codes). That means you have to account for the cost for somebody to
produce those separators. Either you do it in‐house, or outsource it to your
favorite printing company. Yep, another added expense you weren’t
counting on.

• Also as part of your prepping process, you have to manually insert the
separators into the batches so they can “tell” the software where the
beginning of a doc starts and where it ends. (Would you like to do that job
day in and day out?)

• After you scan your 500 pages on the honking scanner, you hope the scanner
doesn’t have a hiccup and misfeed and that your operator doesn’t fall asleep
at the wheel. If so, re‐scan the batch (remember, you are paying for those
page counts).

• Oh, by the way, did you pre‐count all those pages and batches so that you
can be certain the page count the scanner counted matches the page count
the full‐time document prep person counted? Certainly don’t want a
mismatch there! But we digress ….

Let’s look at this challenge from a Distributed, Web‐Based Document Capture
perspective:
• No full‐time employee to prep or scan is required.
• No special facilities or workspace is required to prep the job, much less
actually perform the job.
• Just attach a $200–$400 scanner to the PC, connect to the app via a Web
browser, and away you go.


How about that? No document separators! Does that sound rather appealing to
you? How can you get away with no document separators? Well, first let’s revisit the
reason document separators came about in the first place. It all goes back to that
evil centralized approach that has been shoved down your throat for years. Because
large volumes of documents and papers flooded the mailroom or scan room, the
only way people could keep production volumes up was to prevent the expensive
scanner from stopping. The document separator was inserted between the
documents to tell the software where the beginning and end of a document was.
Remember, it costs the organization dollars when those scanners remain idle, so to
keep them running, the patch code was invented. (OK, it technically even predates
document image scanning and originated in microfilming days).

Fortunately, there is a better mouse trap: Distributed, Web‐Based Capture
software that automatically separates documents based on your classifications. In
Figure 2 below, there is a batch that contains four distinct sets of documents, the
last batch containing a total of 36 pages. (Take special note here that there are no
separator pages, no patch codes, no bar codes to delineate one document or page
from the next. We simply dropped the batch of papers into the scanner and scanned
them. )

Figure 2

Distributed, Web-Based Document Scanning Interface — Mixed Documents.

After we have scanned the documents, we have determined that the last batch
consisting of 36 pages needs to be broken down further into a set of four separate


and distinct documents. How do we accomplish this without rescanning them? Very
simply. All you need to do is highlight the pages you desire to reclassify into separate
documents, click the classify action (Figure 3), and viola! Document separation is
auto‐magically performed.

Figure 3

Distributed, Web-based Document Scanning Interface — Automated Document Separation


Associating Documents Made Simple
with Distributed, Web-Based Capture

Steps for Associating Documents in a
Centralized Scanning Environment
1. Print a large stockpile of available document separators for each location you
desire to perform the scanning operation. This can require thousands of
dollars annually and is a recurring expense (printers, paper, labor, toner) as
separators can only be used so many times before they wear out.
2. Prepare your documents by manually inserting separator pages between
each and every distinct document.
3. Document assembly — organize documents and pages into a suitable batch
size to optimize scanner operators, scanning hardware and software’s overall
performance.
4. Place batches in scanner’s automatic feeder system and press scan.
5. Documents are scanned and sent to the server for further processing.
6. Wait for the server software (usually a server side process, but sometimes it
can be client driven) to examine each page in an attempt to identify the
document separator that you used. By the way, if you use blank pages as a
separator, be prepared to deal with the challenge of documents that
purposely contain blank pages in them (such as the backside of a two‐sided
document) and them mistakenly being identified as separators.
7. You’ll have to insert a post‐QA process into the overall Capture process to
catch those pesky rascals.
8. Correct any errors.
9. De‐prep documents to remove document separators. Plan to reuse the
separators rather than stick them on a shelf or in a box with the rest of your
business documents. Otherwise, every scan job will require new document
separators, and you will be consuming physical storage space to store
meaningless content.

In a Distributed, Web‐Based Capture environment — with features/functions, such
as “virtual separation” — you eliminate eight of the nine steps (and the expenses
associated with them) that Centralized Capture requires.

The bottom line — you don’t have these same document sorting of challenges with
Distributed, Web‐Based Capture that you would with the classic Centralized Capture.
You likely won’t have full‐time employees dedicate to scanning; therefore, it is
reasonable for the document custodian to actually drop the documents into the
scanner — without separators — and let them scan, be virtually classified or
associated via the application. All of this is done through an elegantly designed,
stupid‐simple user interface.


Approaching the Capture problem from a distributed angle makes the process of
scanning, associating, and so on much more simplistic. It eliminates unnecessary
costs and allows the document custodian to maintain a degree of control over the
organization of the documents, rather than leaving the responsibility to a scan
operator 1,000 miles away, who is paid on volume, not necessarily on quality or
accuracy.

Making sense?


Assign Metadata

Honestly, it’s not Rocket Science
You can’t avoid the unavoidable. You need to develop a structured approach that
you can apply to the indexing or labeling of your images or documents. That likely
means that more than one value needs to be assigned to the document you are
scanning.

Take an example in Accounts Payable: One value to describe the invoice, such as an
invoice number, won’t cut it. You need probably five or six values: Invoice number,
invoice date, vendor ID, vendor name, voucher number, and PO number.

You get the idea. That freebie software you get with your scanner doesn’t let you do
that type of sophisticated indexing. (The freebie software usually assigns a cryptic
file name that perhaps you can assign a prefix or suffix to, but that’s a less than
desirable approach when you attempt to scale the operation.) Again, keep in mind
that you have to think of this in terms of scaling the business process and overall
operation. It probably needs to be “repeatable” elsewhere in the company, as you
are contemplating distributing the process.

Next ask yourself, “Am I going to have my staff of A/P or clerical workers manually
label or index each image that gets scanned, or can I take advantage of automating
that process? After all, the vast majority of those values are already contained in my
financial or line‐of‐business system.”

That’s a great question, and, the answer to whether every image must be manually
scanned should always be greeted with a resounding NO, unless proved otherwise.
OK, it can’t always be no, but the majority of values you desire to index the
document with lie elsewhere in your organization in some system, somewhere. (See
sidebar on assigning indexes.)

So now, it raises the question: How do you integrate the values with the images
that you just scanned?

Here is a hypothetical, but realistic Accounts Payable example: A mailroom clerk
receives an invoice from OfficeMax. The OfficeMax invoice references a PO number
that was used to purchase goods.
• The document gets scanned by the mailroom operator and is then routed
to purchasing for the receiving and matching process.
• The purchasing admin staff examines their inbox and notices a new
invoice has arrived from the mailroom. They open the invoice inside the


• Distributed Capture application, take note of the PO number, and then
type that value into the Distributed Capture indexing application PO#
field.
• The application makes a call via a secure Web service to the SAP
application, which the PO was originally generated from.
• SAP returns the value to the Distributed Capture application and also
returns all other relevant indexing data: PO date, PO originator, vendor
number, vendor name — eliminating the need for the SAP purchasing
operator to type any of those values into the distributed capture
application.

With this integrated approach, you have eliminated time‐wasting data entry
cycles and eliminated data entry errors.

Manual Indexing – When it is Really Needed
In our opinion, the only time a user should be asked to intervene with assigning
indexes are under the following three conditions:

1) No database lookup are feasibly possible because the data doesn’t exist
elsewhere in the organization;
2) Documents are not OCR‐able (we’ll talk about that later);
3) At least one value must be manually entered in order to start the
automated look up.

Other than that, users should need only to intervene only to validate the data or
deal with the exceptions.  Period, end of story!

Take it one step further and bar code your documents, and the process becomes
even simpler!

Let’s revisit the previous example but with the addition of a bar code preprinted or
affixed in some form or fashion to the document (Figure 4):
• Upon scanning, the bar code is read, then via an automated, secure Web
service call through the Web browser, that value is compared against the
SAP or Dynamics GP ERP application.

• If the value exists in the application, then the system automatically
retrieves the remaining values and populates the indexing form —
negating the need to manually key those values into the scanning
application.

No human interaction in this example was required except for scanning the
document itself, of course. What if there is a problem, such as the bar code not

being readable, or the ERP system not containing the value, which the lookup is
validating against? We move this code to an exception queue for manual
intervention by an operator.

Introducing this level of automation will dramatically reduce the workload for your
knowledge workers because now they will only be dealing with exceptions, not each
and every incoming invoice!

Figure 4

Distributed, Web-based Document Scanning Interface — Automated Bar-Code Lookups
Distributed, Web-based Document Scanning Interface – automated bar-code lookups

Where do we go from here? Well, if the values are valid, then, via the Web service
call, we can return all the other corresponding values (customer number, invoice
number, PO number, etc.) and completely, yes completely eliminate the need for
manual indexing. All via the convenience of the World Wide Web, safely and
securely!

If you don’t have a bar code, type at least one known value into a field and you
still can perform a database lookup via the web service call, and all the remaining
fields will be are automatically populated. By the way, all this is done without
having to install ODBC tools on the local machine performing the lookups – thus
compromising security; it can all be done in a secured manner via standard HTTP.


Many software vendors will say it can’t be done, that this kind of functionality
requires a thick client. But they would be wrong. Advanced Distributed, Web‐Based
Capture software3 that is available today makes it possible.

3
Such as CAPSYS CAPTURE software from CAPSYS Technologies.


Automate as Much as Possible

Did you really think for a moment that automation isn’t possible in a Distributed,
Web‐Based Document Capture Solution? Traditional Capture industry vendors
would lead you to believe so.

The concept of automated workflow, or what is now referred to as “Business
Process Management,” has been part of the ECM industry for years. However, it has
been locked up in the Content Management repository offerings for over a decade.

The Content Management vendors really did get (and
largely still do) the whole concept of production When looking to
transaction‐based document workflows. The truth is, the automate a business
Capture vendors have largely ignored that enormous process, the analyst
opportunity, thus willingly handing that workflow should begin the
opportunity to the Content Repository vendors. For this analysis as far up the
reason, Capture vendors today have a mediocre workflow food chain in the
offering — if any at all.   process as possible –
even if that means
As previously discussed, the most expensive component going outside the
of the overall ECM solution is the labor to manage the on‐ organization to do so.
ramp or Capture process. Question: Where does the
workflow optimization most logically belong? At the
beginning of the Capture process, immediately post‐capture or after it has been
placed in the CM repository?

We’ll vehemently argue that the workflow process should begin as far up the food
chain as possible. The further upstream, the better chance you have of controlling
the overall process, thus reducing errors and your overall costs.

Thus, workflow optimization belongs within the framework of the Capture
software workflow application — integrated with the CM repository (of course)
and with the line of business application (SAP, Dynamics GP, etc.).

Interface Is Key to Successful Use
While workflow in Document Capture software products is not a new concept, it has
been a rather lightweight offering in terms of functionality. Really, the concept of a
visual design interface, no programming required, robust integration capabilities,
exceptional exception handling, metrics, monitoring, etc. is, shall we say, a rather
foreign concept in our industry.

The traditional software vendors haven’t quite grasped the fact that people relate
better to pictures than code. (Hmm, Capture vendors not understanding the concept


of visuals or pictures that can adequately tell a story, how ironic!) The result is that
they have kept what little workflow capability they might have had embedded in a
scripting language that only a select group of industry consultants can perform on
your behalf. Alternatively, you go to training for a week or two to get indoctrinated
in their proprietary methods and procedures, but only after you have successfully
completed their end‐user and administration product training (another week or
two). Oh well, so be it.

The naysayers have said that workflow capability can’t be embedded in a thin client;
the technology’s not there. It’s there, and it’s been there for years, and there are
vendors that offer it!

Graphical designer tool is a must! Don’t leave home without it
Designing a Capture Process shouldn’t take a staff of programmers or consultants,
but if you look at the legacy vendors, that is exactly what is required.

Do you relish the idea of paying the vendors over $225 an hour for their consulting
time and associated travel and lodging expenses to build a Document Capture
workflow process for you (especially in this economy)? How long do you think it
should take? And, how long does it really take?

The idea of empowering you, the customer, to perform
Are you prepared to the Capture Process flow design and rollout frightens the
spend $0.75–$1.50 in vendors: You are taking away their control, their power
consulting dollars for and their consulting dollars! Thus, the manufacturers will
every dollar you spend push really hard for their team to manage all of the
on Capture software? design, testing and rollout. (Sounds sort of like big
government desiring to have control over all of us
doesn’t it?)

By all means, you need to be properly trained on the product set, understand basic
logic in terms of inputs and outputs, be able to map fields to/from disparate systems
and some other basics, but nothing that is insurmountable. It boils down to time,
money and opportunity cost.

Let’s revisit the better mousetrap: Distributed, Web‐Based Capture. Shouldn’t the
vendors provide you with a graphical interface that allows you to construct your
own Capture process maps, define the various properties of each component or
module, activate them and be done with it? Again, many vendors haven’t quite
grasped this. If they had a graphical interface to construct process flows, you
wouldn’t need to spend $0.75 to $1.50 in consulting dollars for every dollar spent on
the Capture software itself.


It’s interesting to ask — or actually challenge — a Capture software provider to show
its workflow setup process in real time. Give them a genuine business process (being
realistic, of course) and ask that the workflow process be assembled in front of you
as a way to demonstrate the “ease and flexibility” in their Capture
designer/workflow tool. We suspect they will attempt to skirt the issue and revert
back to suggesting a canned demo (just a guess, though).

Programming Not Required, but Available if Needed
No software or SaaS provider can be all things to all people. While a graphical user
interface designer tool is absolutely the correct way to go, no doubt you will run into
a situation where some custom JAVA or ASP coding will be required. As a result, be
sure to investigate whether the architecture is open enough to allow you or your
integrator of choice to develop custom “work‐bots” to automate wherever and
whenever possible using industry‐standard development languages. While the SDK
from the vendor might be proprietary, certainly the approach toward development
shouldn’t be.

The graphical workflow experience and designer tools should provide a visual virtual
assembly‐line approach toward the creation of your Document Capture workflow
construction process. In one Distributed, Web‐Based Capture software system4, the
entire experience is encapsulated from within a single interface (Figure 5) —
beginning with the creation of the initial Capture process (a wizard interface) and
moving to the drag‐and‐drop functionality of the tool palette, to customization of
the configurable work steps, to finally the activate button, which puts your Capture
process right into production (assuming the application didn’t detect any problems
with your process).

4 Such as CAPSYS CAPTURE software from CAPSYS Technologies.


The visual experience looks similar to the illustrations below:

Figure 5

Key Features to Look for:

• Visual Process Designer
• Configurable, Customizable &
Reusable Components
• Field/Index/Meta Data Definitions
types/lengths/masks, etc;
• Exceptional Process Exception
handling;
• Work Queues;
• Automated Tasks;
• Load Balancing, Scalable and
Extensible Server Architecture

Sample of a graphical workflow experience and designer tool user interface, which provides a single
interface for creating and customizing Capture workflow processes. Courtesy of CAPSYS Technologies.


In summary, from a Distributed, Web‐Based Document Capture workflow design
perspective, we believe you should look for the following top 10 features:

1. No complicated software installations should be required for the client to
utilize the Web‐Based Capture application and Web services–based
architecture
2. Rapid Capture Process deployment — measured in hours or days, not weeks
or months
3. A feature rich, graphical user experience that lets the business analyst
designer freely compose the Document Capture process flow, selecting from
a library of predefined work steps and process templates, and using drag‐
and‐drop type features (similar experience to that of Microsoft Office Visio)
4. Error‐checking throughout Capture design process. You should not be able to
put a Capture process into production when it contains missing or invalid
configuration information
5. Database look‐up capability without requiring data source configurations to
be installed locally with the client
6. Ability for the designer to maintain tight controls over each and every step
throughout the process
7. Predictable outcomes for each and every document that is ingested into the
system (whether by scanning, faxing, importing, etc.) without having to go
through extensive testing and debugging procedures
8. Reusable components and process maps. If you build a component or a
process map that is of value elsewhere within your organization, you should
be able to package it up and deploy it elsewhere with very minimal effort.
9. No programming should be required. None! However, if desired, you should
be able build your own custom extensions to the standard product offering.
10. Scalability. Adding additional horsepower to the configuration shouldn’t
require an inordinate amount of time or skill sets. You should be able to
dynamically load‐balance workloads across multiple servers.


Address Exceptions

Exceptions — we guarantee them! There are always exceptions that occur in any
business process. Document Capture processes certainly are not immune to them,
either. Therefore, the Distributed, Web‐Based Document Capture solution must
have built‐in controls to deal with the exceptions.

Let’s take a very basic, hypothetical example to illustrate the point:

Figure 6
A mid‐sized firm sells
widgets on a nationwide
basis and is broken down
into four regions: north,
south, east and west. An
incoming order is received
by the mailroom via U.S.
mail, and that order is
scanned by a mailroom
operator.

After the document is
scanned, the next step in
Order is identified by region; any exceptions must be the business process is to
handled manually. evaluate the origination of
the order by interrogating
Order is identified by region; any exceptions must be the ZIP code so it can be
handled manually. directed to the proper sales
region. In our example, this
is done automatically via
the OCR module (Figure 6).

The exception: Unfortunately, the OCR module can’t read the ZIP code, which
has somehow become illegible. As a result, we have created an exception, and
the software is unable to process the document any further. This is because the
application could not assign one of the four possible outcomes: N, S, E or W.
Hence, the order would be stuck in the Capture process or, worse yet, it may
never be found in the “bit bucket” — resulting in an angry customer.


To deal with these kinds of Figure 7
issues, you need built‐in
exception handling. With a
well‐designed architecture,
exception handling is not an
insurmountable problem!

Every input of a work step in
your Capture process will
always have one of two
possible outcomes: either the
value is true or false. In
Figure 6 above, if the OCR Automated Alert added to exception handling process
failed to identify reduces time and labor. to exception handling process
Automated Alert added
one of the four possible reduces time and labor.
outcomes, N, S, E or W, it
would then automatically route the document to the exception queue for a human
being to manually intervene and resolve the problem. (In
Does the Capture reality, instead of four possible outcomes, our example
product have built‐in would end up with five possible outcomes: N, S, E, W and
exception handling, or Exceptions.) In Figure 7, we even took one extra step and
do you have to code added an automated alert to occur so that an operator is
that functionality into notified by e‐mailed that a document ended up in the
the Capture workflow exception queue.
process yourself?
What is the moral of the story with respect to the issue of
exceptions? Exceptions can exist anywhere throughout
the Capture process; therefore, exception handling must be inherent to the
architecture of the Capture application and all Document Capture processes. This
eliminates any possibility that a loss of critical data can occur.


Alerts and Thresholds
Would you find it acceptable if a purchase order were found lying around in
somebody’s inbox for an extended period of time — beyond your normal set of
business rules?

Figure 8
Probably not.

Well, we expect that you
would not find that acceptable
in an electronic document
scanning system, either!

Here again, you need some
sort of traffic‐cop mechanism
to monitor the individual work
queues that make up your
business process. That traffic
cop should take any and all
necessary and appropriate
steps if the business rules
require immediate action
when a certain threshold
occurs (usually measured in Distributed, Web-based Document Scanning Interface
– configuring Alerts & Thresholds.
time).

Following the same invoice
example described above, if that invoice resides in a work queue for an extended
period of time — we’ll use 24 hours as an example — we want the traffic cop to take
over immediately and automatically route that invoice to another worker so it
receives the appropriate attention.

Setting up alerts and thresholds should be a straightforward task (Figure 8).

Beyond the automatic routing capability upon a threshold occurring, you should also
have the ability to change a value on a field that would result in the triggering of a
new conditional event (such as a route or alert) to transpire.


Archive to Your Repository

Finally, we have arrived — the moment we all have been waiting for! Your
documents have been acquired, the pages have been associated, metadata has been
assigned, you have automated as much as possible and handled the exceptions. Your
content is ready to be published in your Content Management (CM) repository!

Take your pick from any of the CM vendors. You may have even created your own
CM monster yourself. The bottom line — the Capture Software Architecture should
provide the ability for you to easily export or release both the Metadata and the
image information to a repository of your choosing. And of course, deal
appropriately with the exceptions.

Be wary of this
Figure 9
Figure 9
seemingly innocent
expectation because, as
the Capture Software
companies are gobbled
up by the large software
companies, you will
probably see less and
less flexibility with
vendors exporting their
Metadata and images to
SharePoint, FileBound,
etc. Certain vendors may
force you to utilize their
own development
platforms, languages,
An Exception queue withwithAlert function ensuring accuracy
Exception Queue an an Alert Function Ensuring middleware and so on —
Accuracy in the Handling of Every Document
in the handling of every document. thereby increasing the

expertise needed, along
with your costs and time.

Notice in Figure 9 above, there is an exception queue as well as an alert function.
This is because again, you can guarantee that an exception will crop up somewhere
in the process that you didn’t think of, or perhaps a business condition changed ‐‐
thus you need to be able to adequately deal with it. (This can even occur during the
release or export function.) In the example above, we have adequately dealt with
catching those “gotchas” if and when they do occur.


Analytics of the Capture Process

Feedback, Monitor and Figure 10
Control.
Do you like playing
guessing games with
respect to your business,
or do you like to get right
to the bottom line?

Both distributed, Web‐
Based Document Capture
and your expectations
about business analytics
should be viewed in the
same, bottom‐line fashion
(Figure 10).   Know What’s Going On with Your Capture Process In Real-
Time.
Here’s another little secret:
There are some major players in the Document Capture market that have been
around for a very long time, and the concept of using a SQL database to manage the
Capture process from beginning to end is well, shall we say, a foreign concept to
them. If they don’t utilize an SQL database, extracting metric or reporting data may
prove to be a rather challenging task.

Figure 11 Some of the current
vendors use
proprietary file
systems and require
that third‐party ISV
products be
purchased to solve
the reporting
problem. (Without an
SQL database in the
mix, try and see
what’s going on in
real time with the
Administration Features that Let You Move Batches or Documents Capture server,
from One Queue to the Next, or from Worker to Another.
processes or specific
queues. It’s a rather difficult task to achieve.)


Here’s another challenge to consider: As an administrator or custodian of the
business process, at times a situation will arise when you need to personally jump
into the process, working with a batch or a single document, and perhaps reroute
work from one queue to another or simply view the contents of a batch from an
administrator’s viewpoint. You’ll need a specific set of administrator application
tools to do that, and hopefully they, too, are built in to the core functionality of the
Capture application suite (Figure 11).


CAPSYS CAPTURE Software and SaaS:
An Ideal Architecture for Distributed, Web‐Based
Document Capture

To give structure to content (and optimize and simplify business processes), utilizing
Distributed, Web‐Based Capture can begin with CAPSYS CAPTURE software (also
available in a SaaS model). As we’ve addressed the topic of Distributed Capture
throughout this book, here’s a fact: the CAPSYS Technologies solution is truly the
first fully functional Web‐Based Distributed Capture and process management
solution. It can help organizations of all types capture documents electronically from
anywhere paperwork originates.

As a thin client, CAPSYS CAPTURE software is browser‐based and specifically
designed for the Internet. The software gives users a simple, easy‐to‐use enterprise‐
wide tool for capturing and sharing all types of documents at a lower cost than
traditional capture software — without sacrificing capabilities. CAPSYS CAPTURE
software’s client design offers all of the rich production capture features previously
only available in Windows thick client systems. And as we’ve stated earlier, these
systems are costly and difficult to deploy.

Therein lays the big difference. CAPSYS CAPTURE software allows companies to
design their own Capture process graphically, without programming or scripting.
And because time is money, the system is ready to use in days — not months — and
at a fraction of the cost.

Deployed to any desktop browser, CAPSYS CAPTURE software enables documents to
either be scanned or imported from the PC. Documents can then be transmitted
from the client to the server one at a time or in bulk over a secure internet
connection.

CAPSYS CAPTURE software affords users two versions: Companies can readily place
the software into service on their own networks, or they can use it online through a
secure hosted data center. The latter uses the SaaS model, an affordable option
because it eliminates the upfront investment, installation and administration of in‐
house software (as previously discussed). With CAPSYS CAPTURE Online comes an
alternative licensing model: The new software is licensed by server and concurrent
user, not page or click counts. This eliminates the worry of exceeding per‐month
page or click allotments.

All told, CAPSYS CAPTURE supplies many benefits. It eliminates the cost and handling
risks associated with shipping paper documents; increases the speed of transaction‐
oriented business processes; captures and indexes large volumes of documents


securely and efficiently; automates complex capture and indexing tasks; eliminates
manual data entry expenses and errors; imports and processes e‐mail and faxes and
captures documents in their native format to remove costly paper document
scanning; accesses, scans and indexes documents from the familiar Microsoft
Outlook interface; reduces training expenses and increases time to value; and easily
integrates with third‐party Enterprise Content Management content repositories
and legacy systems, assuring a tight fit with corporate processes and systems.

All in for Capturing Effectiveness and Simplicity
Perhaps we’re slightly biased about Distributed, Web‐Based Document Capture, but
can you really blame us? With nearly 20 years in the ECM industry and an industry
knowledge base that, combined, exceeds well over 100 years, we’ve assembled
highly experienced R&D, QA, product management, and leadership teams. Our ECM
backgrounds, business acumen and practical, real world experience include many of
the legacy Capture products in the market that have been industry members as long
as we have. We don’t have decades of old code to support, maintain and attempt to
retrofit or adapt to the changing demands of the way people want to do business.

We saw the emergence of the Internet as the platform (and our competitive
advantage) for the future. The Internet was the foundation we started with. The
legacy vendors can’t make the same claim. From our perspective, we believe the
CAPSYS CAPTURE architecture (Figure 12) will be best suited for the Distributed,
Web‐Based Document Capture market for years to come.

The secret sauce is in our Quick Server Extensions (QSXs). This is what allows our
team of developers, partners and the end‐user customer to extend the functionality
of the core Capture application suite. If you ask us if the product can do “X,” there
might very well be an occasion where we would come back and say to you, “not out
of the box.” However, chances are you can resolve your challenge by quickly
creating a custom QSX built to your specific needs.

It’s not rocket science either. If you know how to code in Visual Studio (pick your
favorite language: VB, VC++, C#, and so forth , CAPSYS gives you the building blocks
via sample code to solve your business problem. We stub out all the necessary code
to plug your custom QSX into the CAPSYS Server framework, so you can concentrate
on the value‐added necessities that may be proprietary to your business process.
Once you’ve compiled your custom QSX, it is made available to the CAPSYS Visual
Process Designer, and you can drop it into your custom Capture workflow process
without any further coding needed! That’s it, as plain and simple as it gets.


Figure 12

CAPSYS CAPTURE Software Architecture Overview


Paul E. Szemplinski is president and co‐founder of CAPSYS
Technologies, LLC. ‐ a leading software company providing
Distributed, Web‐Based Document Capture solutions for the
Enterprise Content Management industry.

For more information, contact CAPSYS at: info@capsystech.com

www.capsystech.com
6755 Earl Drive, Suite 207
Colorado Springs, CO 80918
Telephone: (877) 322‐7797
Inquiries via e‐mail: info@capsystech.com

About CAPSYS
CAPSYS® is the leading developer of Web‐based document capture software. The company’s flagship
product, CAPSYS® CAPTURE, can easily and efficiently capture documents from anywhere at any time
around the world in a safe and secure environment while eliminating the cost and complexity of
supporting additional IT hardware and software environments.

CAPSYS solutions are offered both in the form of Software as a Service (“SaaS”) hosted in the
convenience of our secure data center, or as a traditional “on‐premise” purchase.

To learn more about us or about our proven products, please call 877‐322‐7797 or visit
www.capsyscapture.com.

CAPSYS CAPTURE fully supports the TWAIN standard for image acquisition devices.   For more
information, please visit www.twain.org.

Editor’s Note:
For further information, contact:
Susan Duensing, CBC
Telephone: (847) 639‐8300
e‐mail: susan@rurelevant.com

CAPSYS Logo, CAPSYS CAPTURE, Quick Server Extensions are either trademarks, service marks or registered
trademarks of CAPSYS Technologies, LLC. All other trademarks referenced are the properties of their respective
owners. Unauthorized use of the content may violate copyright laws, trademark laws, the laws of privacy and
publicity, and communications regulations and statutes.


Effective and efficient distributed, web based document capture architecture

More Related Content

Similar to Effective and efficient distributed, web based document capture architecture (20)

More from CAPSYS Technologies (9)

Recently uploaded (20)

Effective and efficient distributed, web based document capture architecture