In October, ABBYY hosted their fifth annual Technology Summit in San Diego. It was an excellent event on digital transformation strategies and ABBYY’s latest technology was laid out. But I got pleasantly distracted by one of the keynotes. During his presentation, Capitalizing on Robotics: Driving Savings with Digital Labor, Mike Gough, senior analyst of KPMG’s Value Architecture Group, explained the different classes of digital labor, which he also referred to as intelligent automation, and how they can help businesses save money.
Gough told us that, according to a McKinsey Global Institute study, 1.1 billion people globally — from a full-time equivalent point of view the equivalent of $15.8 trillion in wages — are associated with technically automatable activities. He also said that most countries, including the U.S., China, Japan, India and most of Western Europe can automate half of their labor forces.
Half of their labor force. Let that sink in.
This started ringing some bells that harkened back to a presentation I attended in Berlin in 2017 hosted by Konica Minolta that highlighted Douglas Coupland, the post-modernist futurist. Work as we know it is changing.
Beyond creating savings in the form of reduced labor costs, digital labor can make financial impacts in other areas. For example, digital labor can help a business capitalize on big data opportunities, and thus create new savings and sales opportunities by utilizing their existing data more effectively.
Gough says there are three different classes of intelligent automation: basic process automation, enhanced process automation and cognitive automation.
Class 1: Basic Process Automation
Basic Process Automation (BPA) solutions leverage rules engines, screen scraping and data capture, and workflow technologies to automate rudimentary “swivel-chair” processes. Gough said that the ideal process candidate is repetitive in nature, has well-defined activities that are easily organized or sequenced, requires little to no tacit knowledge or cognitive assessment, uses relatively structured and consistent data, and can be triggered by some other action.
BPA technology can create savings in a few ways. For one, it can find savings in high-volume business processes. “The higher the volume, the more the costs of automation can be spread over the total number of transaction,” said Gough. “The ‘volume factor’ essentially acts as a multiplier,” he added. Gough also said that since a business can concentrate its activities, “there is a greater likelihood of cost take out,” and that it’s a key to reducing headcount.
BPA can also smooth over some major pain points. For example, humans aren’t the best option for executing certain types of processes because of how inconsistent we are, and because computers are just flat out better at doing some things that we are. But process automation means that every process is executed the right way, every time on the first try.
Class 2: Enhanced Process Automation
Enhanced Process Automation (EPA) technology picks up where BPA drops off. It automates processes that aren’t as structured as those handled by a Class 1 solution, and tend to be more specialized and complex. And considering how bad humans are at completing some simple tasks, it’s not a stretch to say they’re awful at handling complex tasks. These solutions are ideal for complicated processes that require multiple iterations or have complex branching logic.
Typically, an EPA-enabling solution offers up some combination of pattern recognition technology, automated learning capabilities, e-bonding with other popular software platforms, the ability to understand natural language and to consume and leverage unstructured data, and some sort of “out-of-the-box knowledge.” Gough noted that these solutions are typically more expensive than their Class 1 counterparts, but the savings also tend to increase over time because of the built-in learning capabilities.
Gough said that these solutions are more specialized, which can contribute to faster ROI. “The availability of industry- and process-specific starting knowledge reduces payback time, and systems that have an existing library of standard process automations provide a quicker path to automation.”
Class 3: Cognitive Automation
Cognitive Automation (CA) technologies are in a league of their own compared to Classes 1 and 2. They leverage advanced technologies like natural language processing, artificial intelligence, machine learning and data analytics to perceive, infer, gather evidence, hypothesize, reason and interact with human counterparts. Gough noted that CA functions are learned, rather than programmed, and require a lot of time and money to get up and running.
These solutions are the bleeding edge, and are the closest to humans that digital labor can get at the moment. These are the solutions that can read and understand unstructured data–like emails or contracts–think about it, and make a decision. This isn’t your basic robotic process automation.
Automation is one of those conversations that is equally exciting and terrifying. When you discuss this amazing technology, and what it can do, and how it does it — it’s mind blowing. But when you think of the implications of automating half of all the jobs in the world, you start wondering if your job is one of those jobs. But no one can deny that businesses are poised to save a lot of money and create new opportunities on the backs of automation.
The robots are coming.
Latest posts by Patricia Ames (see all)
- Big Tech is Way Too Big to Fail and That’s a Serious Problem - January 16, 2019
- How Blockchain and Bots Are Turning Phishers Into the Hunted - January 2, 2019
- Meaningful Innovation is Messy - December 18, 2018