CloudX Blog

Distributed vs. Centralized Capture in Today's Accounts Payable Process!

Posted by Chris Cosgrove

Aug 3, 2016 4:42:52 PM

Accounts Payable Process Mission Critical!

Today’s Mission:  Review the challenges of distributed vs. centralized capture and how it impacts the Accounts Payable process.

In today’s complex and changing business landscape, many organizations have to grapple with a variety of challenges that beset their Accounts Payable departments.  One of these is the inundation of inbound invoices, especially paper invoices or paper-formatted invoices.  While many electronic initiatives like EDI or e-invoicing have resulted in fewer paper-based invoices, the fact remains that virtually all businesses will have to process some subset of invoices that are paper based or PDF (paper formatted electronic documents).  The challenge of invoice processing is a subject that has been examined at length by us here, but you’ve got some options for how you handle the capture of these documents that really depend on how your business is structured.

For most small to mid-sized businesses, there is typically a centralized AP department and process whereby invoice scanning and capture becomes the norm because it fits hand in glove with your organizational structure.  However, for larger businesses or those who are dispersed over a broader geographic area (ie. a multi location retail format), the challenge of capture becomes more difficult.  Not that the problem can’t be solved, as scanners are inexpensive, reliable, and easy to deploy.  That is of course, if you are of the mindset that we advocate.  That is that invoices need to be converted instantly upon receipt to data.  The data needs to be extracted via intelligent, advanced OCR (not archaic template based OCR certainly, and optimally not by your AP or IT staff) in the cloud, which then gets cleansed dynamically with human operators to bring data levels to near perfect thresholds.


Document On-Ramps


Now in terms of the document on-ramps, there are multiple approaches that would fit this scenario.  For sure, numerous desktop or workgroup scanners could curb the paper by converting it to a digital image.  Alternatively, a monitored, centralized PO box could be amenable unless preventing invoice flow to outlying locations is entirely out of the realm of possibility.  Even leveraging mobile devices for image capture and PDF creation is a possibility in some cases and certainly something that is easily deployable considering the proliferation of smartphones and tablets today.  Anyway it takes, essentially, is the preferred method for routing invoices for capture, but you have isolate the method that is most suitable to your business considering volumes, workflow, capital investment available.  Two other options that would assist with the flow of invoices into AP whether centralized or distributed involve a dedicated, monitored email address which vendors can submit their invoices for processing.  This however assumes that the invoice data contained on the surface level would be enough to auto-identify the key header data which would then enable workflow and processing automation.  In other words, if invoices sent to outlying locations required manipulation of some sort before they are processing by the receiving personnel at the location, then it may be necessary to continue with the aforesaid distributed capture options.  A final, higher-order option is the use of a vendor portal by which vendors could submit their invoices dynamically through a single / bulk upload process into the OCR and validation queue.  So now that we’ve looked at the different cases for distributed capture, it makes sense to turn attention to the centralized capture.

Centralized capture, as we stated above is ideal for small to medium businesses and for large / enterprise level organizations that function in a shared services model.  In this way, you can pool the hardware and processing components needed to streamline the document flow.  We believe a still better way is to process documents dynamically through the cloud, so you don’t have to bring in big ticket OCR technology (which has to be curated and managed extensively and is an expertise that is very nuanced and hard to find in the market) and have your people or hired people manage the data validation efforts.  Besides pooling your personnel and hardware resources, in both approaches to alleviating invoice conversion efforts you’ll want to deploy dedicated email monitoring and a robust vendor portal.  These efforts will directly contribute to reducing the quantity of invoices that have to be physically sorted and scanned.

Regardless of your situation and physical setup, getting invoices converted early in your payables process is the key to unlocking the automation puzzle.  Of course this assumes harnessing OCR and multiple means of automated validation, but the point is that if you’re pushing this method forward, you’re operating in accordance with best practices for handling invoices expeditiously.

Free eBook on  AP Automation vs E-Invoicing:


Read More

Topics: Accounts Payable Process, OCR, Advanced Capture

Using Advanced Data Capture To Speed Process in AP Automation

Posted by Chris Cosgrove

May 24, 2013 1:57:00 PM

Use Intelligent Capture to drive Accounts Payable Automation



The Rolling Stones may have said that ‘Time is on my side’ though for most business professionals time is something that flies by leaving one wondering where it went at the end of the day.  Many people are hard pressed against internal and competitive pressures and as such are looking for ways to make processes more expedient and effective.  One of the advances in document management and processing that has really come a long way in the last 4-5 years is that of Advanced, intelligent Optical Character Recognition technology.  Now OCR is by no means a new technology, dating back over 30 years, but the level of accuracy that OCR can achieve and the business logic that can be applied to newer, more robust OCR engines is very potent and a more recent development in the market.


The net result of OCR then in terms of capture accuracy is in the 95-98% range.  There are many Freeware OCR providers available online and natively within MS platforms, though many of these, to use the term found inside kid’s Riddell football helmets,  are ‘not for competitive play’. 


Certainly there are a slew of robust players in the OCR space including:


ABBYY – International leader in OCR including extensive multi-language support

Paradatec – Highly regarded provider of advanced OCR technologies for semi-structured and unstructured documents.  Used in many high volume transactional processes from mortgage loan application processing to health insurance claims to Explanation of Benefits processing.

Nuance – Widely adopted OCR provider with a cadre of solutions and service offerings, and entry level offerings wrapped around personal multi-function devices for B2C market as well as comprehensive solutions in the B2B space.

Kofax – Another leader in the OCR space with advancements in capture processing to include mobile device capture applications and a highly integrated solution in conjunction with mainstream document/content management applications like Documentum, Filenet, and Sharepoint.


So, while there are many that lead up the mountain there is another element that needs to be discussed in terms of post document capture data conversion.  With 2-5% of on page characters being categorized as ‘low-confidence’ or mis-characterized, it is necessary for a dynamic data audit to occur to make the data fully use-able.   There are a couple of approaches to this and some BPO providers go so far as to have multiple OCR engines process the same images and then overlay the results narrowing the accuracy gap by better predicting what the OCR believes to be the accurate character.  While this is a step in the right direction, albeit costly as many OCR providers charge their users on a per click (image processed) basis, it does not ensure 100% accuracy.  With that said, we believe that the best way to achieve high levels of accuracy is by overlaying human intelligence for the low confidence characters.

Essentially then, by having Data Entry operators exclusively focus to low confidence characters on-screen, you can achieve extremely high processing rates because the operator is focused to looking at data in context.  This enables high production capability because the volume of data entry is greatly diminished, and the nature of the entry is essentially validation and correction.  By synthesizing and harmonizing the advanced technology and the process, you can achieve much higher levels of productivity than leaving either approach to its own devices.


Processes ripe for improvement with these technologies:


Accounts Payable Invoice Processing / Accounts Payable Automation

Healthcare Explanation of Benefits (EOB) Processing

Insurance Claim Processing

Mortgage Loan Application Processing

Sales Order Processing

Supplier Statement Processing

Human Resources Employee Onboarding


The key take away for any burdened process is really that in today’s networked economy, the tools you need to truly transform and optimize business processes are no longer out of reach.  Instead, due to cloud based document management and processing, and by adopting some of the best practices associated with Document Process Outsourcing(DPO), it’s possible to make obsolete and inefficient processes to be highly effective.


To learn more about DPO best practices check out our Resource Library!

Read More

Topics: Accounts Payable Automation, Improve Accounts Payable, Invoice Processing, OCR, Advanced Capture

Subscribe to Email Updates

Posts by Topic

see all