If ever a job cried out to be automated, it was the one I landed fresh out of college. I was working at a nonprofit and all set to change the world. Then my manager pointed to a box of documents and told me to create a database. Tedious weeks later, bleary-eyed, I plopped the box back on her desk and asked for another assignment -- any assignment.
Mine was the sort of mind-numbing experience that's inspired many a junior office worker to pine for a day when automation can take over. But surprisingly, scanning and sorting documents at scale has been something of a white whale in the records management industry. There are plenty of industrial digital scanners, but they rely on humans to do the sorting.
Backed by Steve Wozniak and Google Ventures, a company called Ripcord has been using AI and advanced automation to solve the problem. In an age where data has never been more important, digitization offers a way to bypass the tedium and inefficiency of manual entry. Throw a robot in the mix and a stack of paper representing reams of unstructured data becomes discoverable and manageable information.
"Ripcord has built a robotic digitization and content enrichment pipeline with the intent purpose of rapidly creating valuable data for our customer's. Every subsystem on this new machine has been optimized for speed, autonomy, quality and security, which will enable us to deliver this service at an unprecedented scale," said Kevin Hall, Ripcord Chief Technology Officer and co-founder.
But building a robot to perform data entry from physical media is no easy task. In spite of a century of technological development, automatic paper handling is precarious, as deftly explored by Joshua Rothman in the New Yorker. Ripcord's bots are built to handle everything from ledger pages to tiny business cards, and to know the difference between them so it can sort the printed data accordingly. The last part is possible thanks to huge leaps forward in text recognition, which has gotten remarkably accurate in a short span of time.
Ripcord's newest robots, which are smaller than early generation predecessors, combines robotics, computer vision, and AI in a nifty technology stack. Papers go in, and the robot kicks into gear, using more than 50 sensors, including sheet detectors, cameras, and laser measurement, to scan and digitize the media, then place information in the cloud, where it can be sifted and analyzed.
So far the company has raised $65 million on a pitch to disrupt the $25 billion record management industry. Part of the selling point is security. The information I was manually entering into the database at the nonprofit, for example, dealt with the mental health issues in a specific patient population. In the wrong hands, that becomes a massive security breach.