The Forgotten Element in Your Big Data Strategy
By HK Bain, CEO, Digitech Systems
In all of the discussions of ‘big data,’ it’s easy to get overwhelmed by the incredible figures about the growth of digital data. As a result, many organizations overlook the critical information that is stored in paper files. It can be thought of as the forgotten element in our big data strategies. Neglecting paper-based information when making business decisions could significantly limit our ability to make smart choices, but a major shift in technology development to integrate Artificial Intelligence (AI) algorithms into business applications will improve our ability to manage all of our big data—whether paper or electronic.
The Forgotten Element
Paper documents still represent a significant source of business-critical information that can improve decision-making in at least a couple of significant ways.
First, paper files can lend an important historical perspective to data that is being mined for trends. For example, let’s say an individual was diagnosed more than ten years ago with a chronic medical condition that requires frequent blood tests. If today’s healthcare provider notices only a slight elevation in important blood levels by viewing the two to three years of blood tests stored in the patients’ Electronic Medical Record (EMR), they may not see a significant change that has developed slowly over the ten year period. The patient may not receive needed treatment, because the larger but slower moving problem is missed without the historical perspective.
Second, paper-based information is still the norm for many parts of your organization. I know our knee-jerk response is to deny this, but assessments frequently reveal that individuals and departments are hanging onto a large number of paper files. For example, the Institute for Finance and Management explains that 70-80 percent of invoices in the United States are still paper. Without the ability to cull data from these paper files, we risk making decisions from an incomplete perspective. In fact, if that decision needs to weigh certain financial data, we could be off by a huge margin if we ignore 70-80 percent of the data simply because it’s not accounted for in our big data systems today.
The ability of artificial intelligence to more quickly and accurately sort information can play a huge roll in streamlining processes
What’s New with Paper?
The ease and accuracy of extracting critical business data from paper got a boost last fall when companies began integrating artificial intelligence into capture applications. Document capture, or scanning, applications work directly with scanning hardware to create a digital image of a paper document and then to allow operators to apply indexes or informational tags to the document that make it easy for users to locate it using keyword search. Traditional scanning applications require users to sort documents into types manually and then to hand-key the indexes while viewing the paper original or the scanned image. In some cases, applications use Optical Character Recognition (OCR) to then find certain words or phrases within a document to populate index fields.
Artificial intelligence simplifies this process by eliminating manual steps and generating more accurate results than are available through simple OCR. Machine learning describes a category of artificial intelligence in which software algorithms learn from large data sets without relying on the more traditional conventions of rules-based development. This allows the computer to explore an exponential number of options before selecting the right answer. AI algorithms enable capture solutions to receive many different types of documents at once and to sort them into categories based on patterns—without pre-programming those categories into the application, dramatically reducing the setup time required to prepare paper documents for scanning. With AI, you can simply drop the entire stack onto the scanner, and let the software figure out which pages are invoices versus letters, versus forms, etc. Artificial intelligence engines are also improving the accuracy of data extracted from the document to create index values, because they look at many more data points before making a decision.
The ability of artificial intelligence to more quickly and accurately sort information can play a huge roll in streamlining processes as demonstrated by MSI Mold Builders. Headquartered in Cedar Rapids, IA, Mold Builders designs and makes custom manufacturing molds. They receive about 1200 invoices monthly, and were falling behind in their paper-based processes. To help automate AP processes, they implemented an ECM system in 2015, which included an artificial intelligence engine to classify invoices by type and extract important data for their accounting system. Thanks to the AI assist, they have shortened invoice processing time by more than 45 minutes per invoice and reduced the cost per invoice by almost $50! Jason Sojka, the Network and Computer Systems Manager said, “The PaperVision Forms Magic technology has revolutionized our AP processes! We’ve been able to cut our invoice processing times by more than 75 percent! We’re saving the organization money and improving relationships with our business partners.” Mold Builders saves about $676,000 every year thanks to the AI assist in the scanning technology they chose to handle their data problem.
According to Research and Markets, the market opportunity for AI systems for enterprise applications will grow from $202.5 million in 2015 to $11.1 billion by 2024, a CAGR (Compound Annual Growth Rate) of more than 56 percent! That makes AI one of the fastest growing market segments overall, and it will be a great place for IT to have expertise over the coming decade. AI algorithms are already improving the capabilities of everything from big data software solutions to manufacturing robots, so chances are there is already something out there that could help your organization to save time or money. You just need to find it!
Robotic Refactoring the Workplace
Philip D. Heermann, Senior Manager, Sandia National Laboratories
The Holy Grail of Customer Itinerary: Open AI Ecosystem and Contextual insights
Eric Saint Marc, VP-IT, Palms Casino Resort
Why Your Next Insurance Claims Processor Could be a Robot
Rod Dunlap, Director, Alsbridge
Building an AI Based Machine Learning for Global Economics
Alexander Fleiss, CIO & CEO, Rebellion Research Partners LP