Artificial intelligence-powered software is increasingly deployed in the transition to an automated workplace. But far from rushing towards a dystopian future where humans cower under a robot’s boot, automated systems are expected to free us up from the most mundane tasks.
Mass job losses? Not going to happen in the new workplace, the developers of automated systems assure us. Instead, humans will do what we arguably do best – creative problem solving.
A number of companies have become well known in this corner of the enterprise software market, with robotics process automation (RPA) start-ups such as UiPath, Blue Prism and Automation Anywhere all making headlines this year.
The latter, whose machine learning system can complete repetitive tasks including filling out forms, attracted $290 million in new funding in late NovemberOpens a new window , valuing it at a hearty $6.8 billion. Another start-up, UiPath,Opens a new window raised $568 million in April for a valuation of $7 billion. And in June, Amazon launched TextractOpens a new window , a cloud-based managed service that uses machine learning and character recognition to extract data from documents.
GartnerOpens a new window researchers say robotics process automation is the enterprise software market’s fastest-growing segmentOpens a new window , with revenue estimates of $1.3 billion this year, up 63% from $846 million in 2018.
The biggest adopters of automated systems, the research group finds, are companies with so-called legacy systems, such as banks, insurers and utilities. At the same time, large software players like Microsoft, IBM and SAP are taking note and buying up or partnering with robotic providers as the machinery of efficiency moves ever onwards.
The sunset of data entry
Enter a Czech outfit called RossumOpens a new window , whose professed mission is to rid the world of manual data entry. It says that every day across the world, the equivalent of about 100 human lifetimes are spent on manually entering invoice dataOpens a new window .
Several seed-funding rounds have snagged $4.5 million in padding for the startup to launch its AI-based data entry tool to speed the processing of invoices and other forms.
Rossum claims its product can extract data from documents six times faster than a human worker and is 95% accurate. It can also save companies up to 80% of the personnel costs of manual processing — not to mention the savings on recruitment when workers leave or are replaced, their minds numbed from hours spent hunched over mundane tasks.
While not strictly a provider of RPA, which Rossum says is better suited to structured data, the Prague-based company offers a complementary service transforming unstructured information into a structured form.
“I saw it as a very complementary offering to UiPath and what the robotic space was doing,†Tobias Rataj, head of partnerships and alliances at Rossum and formerly a vice president at UiPath, tells SiftedOpens a new window .
Cognitive data capture
Traditional optical character recognition extraction software relies on specific rules and templates for the documents it inputs, reading each character separately. The creation of templates and rules, as well as developer maintenance, can be time-consuming and costly.
Rossum’s potential edge lies in its use of cognitive data capture, which instructs computers to recognize documents as people do. Its cloud-based software uses deep learning to scan documents, regardless of formatting or style. If any data field is not recognized, feedback is sought from a human employee, with the system learning and improving from each feedback interaction.
While primarily involved in processing invoices and delivery notes, potential use cases include accounting, logistics and insurance.
Cognitive data capture has been used for several years by IBM, which launched a product to help businesses extract more information from enterprise documents in 2015. Another player using the technology to turn documents into data is Kofax, ranked sixth in the RPA market by Gartner.
“Technology should make data entry easier and cheaper but businesses have become too reliant on using old systems that no longer meet their needs,†says Rossum’s co-founder Tomas Gogar.
The company, he says, “solves these problems without complicated, clunky integrations; without teams of developers; and without high costs.â€