Original article can be found here (source): Artificial Intelligence on Medium
Cognitive Software Group — Intelligent Process Automation “IPA”
Eliminating non-compliance with regulations, and prospering.
In the past five years we have seen stunning growth in a software market segment known as Robotic Process Automation. “RPA” is a relatively new way of achieving business process automation utilising software robots (bots). Basically, the robotic software observes and replaces human knowledge worker interaction with computer programs, lowering the cost of performing the process and freeing up the knowledge worker to do more useful things.
RPA has experienced a boom in recent years. According to Gartner, it grew 63.1% in 2018 to $846 million and was expected to reach $1.3 billion in 2019, making it the fastest-growing segment of the global enterprise software market.
It sounds great, but it’s not plain sailing.
Knowledge workers complain that it is often used as a means of making workers redundant. In The Guardian on March 3, award winning writer Ellen Wengert describes her experience with this relatively new technology saying “I didn’t exactly love my dead-end data entry job, but wasn’t thrilled to be losing it to robotic automation either”.
Writing in the Harvard Business Review in mid-2018, Thomas H. Davenport and David Brain suggested “before automating your company’s processes, find ways to improve them”.
“To be clear, however, the match between RPA and business processes isn’t a perfect one if the goal is to redesign or improve the process rather than to automate its current state. As Andrew Spanyi, the author of four books on process management, put it to us by email: “RPA does not redesign anything. It doesn’t ask whether we need to do this activity at all. It operates at the task level and not the end-to-end process level.””
“Many RPA implementations support the “as-is” process, with no improvement or examination of the current process steps that are automated. As a result, they may achieve modest savings, but in many cases, they will miss out on opportunities to dramatically improve process outcomes, quality, costs, and cycle times.”
Recently, RPA vendors have added Machine Learning techniques to their platforms claiming that this makes them intelligent. Given that Machine Learning is a pattern matching technique it is not going to provide “intelligence”; you need a Semantic Graph database for computer intelligence. Nevertheless, Machine Learning is a useful technique that may highlight patterns in the process that indicate poor process in achieving optimal outcomes, e.g. a process that causes anxiety to customers.
Writing in Medium, a Singaporean RPA start-up, CFB Bots complains that Intelligent Process Automation “IPA” is a misleading term implying that RPA can be made intelligent. Quoting the IEEE, they argue that, “put simply, you can think of RPA as a software robot that mimics human actions, whereas AI is concerned with the simulation of human intelligence by machines”.
They continue “Before we go into the differences between the two technologies, it is important to realise that RPA and AI are nothing but different ends of a continuum known as IA.
In this respect they say “IA” is “Intelligent Automation”, or Intelligent Process Automation.
A somewhat more robust argument is put forward by Blogger Horses for Sources, opining that “The RPA that “died” is the poorly-defined “RPA” that got hyped up to create hockey-stick growth excitement for investors. It wasn’t defined correctly, was a mash-up of desktop automation with pure-RPA (unattended back office) and all the deals that got signed were “attended” so weren’t even “robotic”.”
More defining of the future “intelligent RPA”, SAP Communications Director Susan Galer writes ““In some industries, intelligent robotic process automation just might become essential to survival. IDC predicted that by 2023, 60 percent of G2000 manufacturers will address growing industry talent shortages with significant investments in intelligent RPA. These analysts wrote that “the integration of optical character recognition (OCR), natural language processing (NLP), and machine learning in AI-infused RPA opens up opportunities for data collection, workflow, and operational/tactical decision making.”
As the only Australian company and one of the few world-wide offering a full-service, build-your-own cognitive computing workbench, we agree with all of the points made in the articles quoted above, especially the latter. We typically propose a combination of Semantic Graph, OCR, Machine Learning (including Deep Learning), and NLP, for a true AI solution. Without all of these techniques the benefits may be attractive, but they will be incremental.
What an opportunity AI really is, and it isn’t that hard.
In recent times we have taken a look at the most infamous business processes in Australian history, the processes (or lack of them) that gave rise to the $70 million Royal Commission into the Australian Banking Industry, established on 14 December 2017 by the Australian government. The Commissioner’s final report made 76 recommendations relating to institutions and individuals that have engaged in dishonest misconduct and charged the regulators with the responsibility for taking action. The Chairman and CEO of one of the leading Banks received particularly strong criticism leading to the resignation of both.
The ABC reports that “Billions of dollars of stolen money is being returned to customers, multiple court cases are underway and only one of the bosses of the big four banks has passed the milestone of two years in the top job. About $1.5 billion in remediation has been returned to customers, and the total is expected to hit $10 billion.”
A cursory look at the environment created for investment advice, funds transfer, lending and so on suggests that there has been catastrophic failure on the part of governments and industry to protect the billions of dollars’ worth of assets of Australian consumers against non-compliant, unconscionable, and criminal conduct. The financial services industry complains about over-regulation while it relies on 1980’s technologies to support 1980’s business processes. The regulators complain about industry lobbying and the lack of criminal convictions. The Government puts the Royal Commission report on a high shelf out of reach without a ladder. Consumers, who knew it was all crooked throughout, put it all in the too-hard basket until the next election.
On bringing down his report, Commissioner Hayne refused to shake hands with the Treasurer of Australia.
As observers to this and having looked under the covers ourselves, the financial institutions of Australia have clear choices about the new Artificial Intelligence technologies. If they embrace the opportunities, they can massively reduce costs, deploy their key staff in more strategic activity, develop and implement a customer experience strategy informed by Intelligent Process Automation, and eliminate non-compliant practices.
While we don’t necessarily believe the institutions should interrupt the remediation processes by introducing new AI technologies to them, the remediation activity offers a minimum-cost opportunity to use the lessons learned from the old to build new automated processes based on cognitive computing in parallel to remediation of the old manual processes.
If only one of them embraces the opportunities, they are likely turn the rest of the industry into a mere rump.
If not, history may judge all of them to have ignored Ms. Galer’s insights above, to their peril.
Author: Mark Bradley is the Founder and Sales chief of Cognitive Software Group, the leading cognitive computing company in Australia. www.cognitivesoftware.com