On Earth Day 2020, much of the world is staying home because of the COVID-19 pandemic and getting accustomed to the benefits and challenges of working remotely. However, many business processes, from inventory to invoicing, still involve paper records that are not remotely accessible. Ripcord Inc. is applying lessons from industrial automation to the digitization of documents and processes, helping organizations move away from decades’ worth of paper usage.

“I have a background in industrial automation,” said Kevin Hall, co-founder and chief technology officer at Ripcord. “It was eye-opening to learn how prevalent paper documents are for transmitting and storing data. We started to look at why that is and found millions of boxes of paper stored in warehouses.”

Hardcopy records are kept in a wide variety of files, folders, binders, and paper types and sizes. Any bulk scanning requires human preparation of the documents, which can be slow and labor-intensive.

Ripcord takes a factory approach to content

Ripcord scans more than 1 billion documents per year in Hayward, Calif. It uses robots to sort pages, maintain privacy, and remove staples 10 times faster than a human. People are involved in fewer steps than in a typical processing facility.

“We used techniques from factory automation for preparing content,” Hall told The Robot Report. “Conventional technology was not able to meet the scale of the demand. We designed and use our own machines and use state-of-the art hardware and software in our central facility.”

With machine vision and artificial intelligence, Ripcord creates digital twins of documents. “Variability and scale are challenges,” said Hall. “Both the physical content and data in images are highly unstructured. We built a solution to tackle both, starting with a core workcell that’s an integrated piece of automation.”

“It includes some off-the-shelf and custom components,” he said. “The speed and flexibility of a SCARA robot arm comes into play, and we use numerous sensors and industrial linear actuators. It uses computer vision and machine learning to learn what to do.”

Robots handle multiple inputs

Most pick-and-place operations involve a set number of items, which might be large in e-commerce. Unstructured inputs can be found in another industry that’s starting to adopt automation for dealing with paper, among other materials — recycling.

“This is unique to only a few industries, where robots have to deal with such inputs and improve over time,” Hall said. “There are large and small documents, different types of content, and a wide variety of fasteners and binders.”

“We use vision systems to identify and classify items and make automated decisions,” he said. “We’ve heard of offices trying to go paperless for years, and we’re now helping companies to go digital.”

“Other vendors have tried this internally or told us that they were looking for a better way to digitize paper records,” said Hall. “With manual sorting, people get trained to look at the top left of a packet for staples. If they miss them, the scanner would find them, often resulting in ripped pages. We found in our benchmarking and testing less paper being ripped.”

“By helping companies become less reliant on paper, the hundreds of millions of tons now stored in warehouses could feed the recycling pipeline, reducing the need for new paper,” Hall said. “We encourage customers to recycle, and the quality of data we turn their documents into is a major incentive.”

An ‘end-to-end pipeline’ for data management

“On the software side, Ripcord is extracting data from the images themselves,” Hall said. “We’re a driving force for organizations to get better organized and more secure. We go through continuous security audits for customers such as large banks and major corporations. Many legacy digital repositories don’t have good controls.”

“We approached digitization as a single problem and built an end-to-end pipeline, from documents to data and then business processes,” he explained. “There were a lot of challenges that we’ve solved with a software-first approach. Our systems look at where content was shipped from, what’s on the manifest, and what was in the actual documents.”

“On the data-extraction side, we continue to be intrigued by possibilities from customers,” added Hall. “Everybody’s needs are a little different. We offer full-service digitization and extract classification data, key-value pairs, and, by default, OCR [optical character recognition] to make it searchable.”

“Our Canopy cloud-based platform enables customers to search and interact with documents,” he said. “It’s common to connect data to other line-of-business processes or RPA [robotic process automation]. There’s a wide spread of where data can go once it’s extracted as part of digital transformation.”

“We’re one of the best on ramps into RPA workflows,” Hall noted. “Think about a document entering a business-process workflow. If it arrives at the best quality with structured metadata, it’s easier to make a decision of where that document should go.”

Serving on an industrial scale

Ripcord, which raised $45 million in Series B funding in February, has major customers including Coca-Cola, BP, Chevron, and UCLA.

“We’re helping Coca-Cola with invoice processing for its North American deliveries, which is still heavily paper-based,” said Hall. “This leads to structured data for its ERP system for billing services. Ripcord is a critical part of that pipeline, allowing the company to consolidate vendors.”

“We serve industries including banking, financial services, and human resources,” he said. “We’ve also done a fair share of healthcare use cases. It takes a little longer to get operational compliance with regulations like HIPAA [the Health Insurance Portability and Accountability Act], and we have serviced pharmaceutical companies.”

What’s the optimal amount of records to take advantage of Ripcord’s robots and services? “The larger the company, the more paper it’s moving and storing every day,” Hall replied. “When we came out of stealth a few years ago, we were messaged by people with boxes in their garages. But smaller projects for us start at 1,000 boxes.”

“Most customers have up to 100,000 boxes. At 2,500 sheets per banker’s box, that’s 250 million sheets of paper,” he continued. “Smaller customers have north of 1 million sheets of paper. We’re also working with federal, state, and local governments, which are inundated with paper for different functions.”

Ripcord continues to innovate

Ripcord has plenty of business, but there are always more types of documents to digitize. The company is planning for regional facilities in North America, and it is actively working on an expansion in Japan.

“We keep pushing the technology forward,” said Hall. “We’re working to automate the scanning of additional content types and deal with different types of folders, clips, and portfolios.”

“We’re also looking at improving vision systems for the document-conversion process,” he said. “We’re looking fo ways to balance automation with data extraction, and we’re looking at how vision systems can change how we handle things like handwriting.”

As the novel coronavirus crisis continues, Ripcord has sent much of its staff home and is helping companies continue to work without centralized paper processes. “We play a critical role in the supply chain of a number of our customers,” Hall sad. “It’s more critical than ever that companies can efficiently connect with invoice or payment processes or shipping manifests remotely.”

“In our own facility, we’re looking at where we can add automation. There are plenty of opportunities for robotics, including fleets of machines for material handling. We’re looking at being the factory of the future.”