Document Data Extraction Gains Automated Smarts for a Range of Use Cases

Digital transformation at the local Department of Motor Vehicles (DMV) might sound a bit like colonizing Mars. Still, a range of innovations in document automation is changing that mentality and breathing fresh life into informational upgrades ranging from government agencies to commercial enterprises. 

Speaking with PYMNTS’ Karen Webster for our J.P. Morgan Chase Merchant Series: Global Innovators in Payments, Ulf Persson, CEO of ABBYY, said advances like intelligent document processing (IDP) and intelligent process automation (IPA) are advancing to the point where companies can realize new efficiencies in document scanning and add value in the process. 

A pioneer in optical character recognition (OCR) technology with its FineReader tool introduced in 1993 to extract data and text from static documents like PDFs and make it editable, he said ABBYY’s focus now is digital usability for business and government. 

ABBYY was doing a strong business in newer information capture applications when robotic process automation (RPA) burst on the scene around 2017, disrupting the space. 

Persson said ABBYY re-engineered its products to fill gaps RPA wasn’t very good at, like extracting business information from document-centric formats. 

That created a new opportunity for “those of us who were good at document understanding or intelligent document processing [and] found a way to expand our addressable market by a factor of many times,” he said, making basic image capture “stale” but bringing fast growth to IPA and IDP. 

Sectors reliant on official source documents are where demand for IPA and IDP is gaining ground, he said, noting, “In our case, financial services, banking, governmental services, transport, and logistics, have probably been the largest, as well as healthcare to some extent.” 

See alsoToday’s CFO Needs to Understand Companies’ Hidden Costs 

Freeing Trapped Data 

Asked about the “intelligence” yielded by intelligent document processing and intelligent process automation, he said it “comes from being able to not only find but also extract information that is otherwise embedded in a document and make that available for all kinds of business processes.” 

Not only is it valuable in remote onboarding, but it also touches on insurance claims and other processes “where that data is generally necessary, but a dumb process would not be able to extract and use it. It’s effectively about making data usable and valuable in business processes.” 

In action, this brings to life digital transformation at DMVs — ABBYY is now working with California and others on this — and for banks, stores, and anywhere identity is vital. 

Citing the California DMV use case, he said, “That’s now all automated. You can do it from home with the help of your mobile. From a user perspective, it’s fundamentally transformed, but also from the point of view of how they do business” in that staff is freed up for higher-value tasks. 

“It’s not about replacing people,” he said, “it’s replacing mundane tasks so you can up-skill people to do more of what they’re good at, and what the machines are not so good at doing.” 

Documents are still the source, but using IDP and IPA that incorporates likeness and liveness detection in a platform environment, he said, “The whole purpose is that you package it all into one solution. You don’t have to do every one of these steps in different windows, as it were.” 

Getting on the Happy Path 

Pressed about the next phase of innovation where the paper or PDF source document comes out of the equation entirely, Persson said that’s on the horizon but still some way off. 

“In the foreseeable future, it doesn’t change very much whether this is a digital document, a PDF document, or where it comes from,” he said. “If it’s in a file or an email or somewhere else it still has to be found. It still has to be extracted.” 

As to the business demand for this type of automation, he sees process mining as the closer frontier because businesses underestimate the difficulties in automating information. 

“You think you know the process, but when you map and mine and analyze, it turns out in nine cases out of 10 that you don’t. It goes all kinds of ways that you don’t think of. If you think about the so-called ‘happy path,’ that’s maybe 20% or 30%” of outcomes now,” he said. 

“However, working with partners, he said ABBYY can define processes that can be automated “and help them measure the success of that automation almost in monetary terms.” 

Regarding document understanding, “we are typically either the lead or we’re part of larger transformation initiatives where we do the document understanding bit. We are then oftentimes invited directly by the enterprise” to provide expertise in complex documents. 

Saying it’s tough to make projections in this economy, Persson said the sector is still in its adolescence and that “enterprises, banks, and governments want to partner with someone they believe can be an innovation partner together with them,” focusing on delivering enterprise-grade quality.