The mix of AI, RPA is turning business process automation smart, IT News, ET CIO
Developing artificial intelligence for business process management is not an easy task. Many companies add AI to processes by creating or buying single-task bots such as NLP systems or vision detection tools and adding them to processes using traditional non-AI methods.
However, human intelligence is still required to filter out processes, break systems down into a single coherent process, change processes as the business progresses, and identify and fix problems.
According to McKinsey, AI, machine learning, and related technologies are now invading this area via robotic process automation (RPA). This combination of AI and RPA leads to intelligent process automation (IPA). In addition to RPA and machine learning algorithms, IPA also includes process management software, natural language processing and generation, and cognitive agents or “bots”.
According to McKinsey, IPA can improve efficiency by up to 20 to 35 percent, reduce process time by 50 to 60 percent, and increase ROI in the three-digit percentage range. It is early days, however, as most companies are in the early stages of development, using individual AI pieces, and rarely combining them into a complete end-to-end automated process.
“There aren’t any use cases that go all the way,” says Gartner analyst Moutusi Sau about the introduction of RPA in the financial services industry. “There were some chatbot engines and AI decision-making tools, but you can’t build on one particular solution. Banks want to do more than one thing. “
The humble bot
The German ZF Group, an automotive supplier that started applying information to its business processes a little over a year ago and started creating a bot to answer the repetitive questions.
“In our corporate communications department, we have a lot of repetitive work,” says Andreas Bauer, the company’s IT manager. “We have a lot of emails in our inbox with a lot of repetitive questions.”
However, once most of the steps in a business process are automated, another level of intelligence can be applied. When selecting the providers for its bots, the company kept an eye on this future.
“We are moving in the direction of automating the entire process chain,” says Bauer. “We weren’t just looking for a bot. We were looking for an orchestration and integration platform where we could easily adopt these technologies and combine them with intelligence.”
While automated integration and orchestration is the end goal, the company also wanted a platform with built-in checks and balances. “There was a fear that something would go crazy and we couldn’t control it,” he says. “You have to be careful, you have to keep an eye on the technology. It’s not like the technology is self-sustaining. You have to make an effort.”
The ZF Group chose Vizru, a bot platform that offers management, governance and language support layers among the bots. It is known as the Stateful Network for AI (SNAP) process and it stops a bot when it exhibits abnormal behavior. According to Sizru, the SNAP layer can also flag or pause a transaction if there are compliance violations or confidential data is inappropriately exchanged between processes.
Another option is to add smart decision points to a traditionally automated business process.
This is exactly what American Fidelity Assurance, an insurance policy provider, does. One challenge for the company was to automatically route the many emails it received every day to the right destination. In the past, someone decided where every email should go.
“Is there a way to get advanced machine learning to learn from previous data and decisions and make the same decision that a human would make?” asks Shane Jason Mock, vice president of research and development for the company, who was inspired by a tour of Amazon.
American Fidelity turned to UiPath, an enterprise RPA provider, and the DataRobot AI platform to add information to their processes.
“In the new email process, we combined the RPA component with the machine learning component, and the combination of the two decides where the email needs to go,” he says.
In many cases, traditional RPA approaches encounter a decision point that is too complex to be easily automated. The company also intends to use AI for process mining to automate process discovery rather than business analysts figuring out what’s going on in the company.
The traditional approach to business process management involves business analysts speaking to managers and employees, performing audits, and then creating diagrams that illustrate the various business processes in the organization.
“Many of the customer engagements we work on have a process workflow on the wall,” says Sumeet Vij, director of the strategic innovation group at Booz Allen Hamilton. “But do things really happen that way? You will find that the way things actually happen is different and the bottlenecks are different. Using machine learning for process mining, people can get an idea of how things are actually going. “
As the business evolves, these tools can update the processes and also detect abnormal behavior in real time.
One company that already has an intelligent process mining system in place is Chart Industries, a manufacturing company for the power industry headquartered in Ball Ground, Georgia.
Chart Industries, a manufacturing company in the energy industry, had problems a few years ago. The company’s share price fell, and the top managers were replaced, and new leadership wanted to make changes. For example, Chart had three main departments, and although they shared a single ERP system from Oracle and JD Edwards, there were several back offices that handled accounts payable, accounts receivable, and other back office tasks – each with their own processes and procedures.
“We found that our customers were effectively benefiting from paying us later than they should,” said Bryan Turner, executive vice president of IT at Chart.
There were other ways to influence cash flow as well. In some cases, for example, the company may take advantage of discounts for paying providers within a certain time period. In other cases, it might be beneficial to hold on to cash longer. The benefits of better efficiency can run into the millions here, says Turner.
Chart reached out to Celonis, a process mining company, to uncover such opportunities.
“We ran it on some custom systems today. As long as it has a database, transactions and timestamps, you can enter them into Celonis, ”says Turner. “A big part of the heavy lifting was moving data between our organization and the SaaS application or the Celonis Amazon backend.”
The business process can be viewed in the form of diagrams such as Visio diagrams, and managers can drill down into the process down to the level of each transaction.
“Only in one late payment example have we seen $ 240,000 in annual savings,” said Turner. “The software has paid off several times and we continue to see that the cost option definitely works for both our suppliers and our customers.”
Business process analysis
Seann Gardiner, senior vice president of business development at DataRobot, a provider of AI platforms, says some of the most progressive companies have enough business process data to now see the big picture, analyze and predict.
“They’re taking the exhaust from the RPA process and trying to capture that and learn from it and make those processes smarter,” he says. “I wouldn’t say we see it very broadly in organizations, but we are starting to see it.”
If a company has a strong focus on process-level automation and can turn that data off, they might be ready, he adds. “But you have to have leaders who believe in automation, an AI-first mentality, and can make the organizational changes you need.”
Fortune 5000 companies are ready, he says, and have processes in which they can use a combination of AI and RPA, he says. “The question is, do they want to put in the work to make these big changes to the organization?”
The article first appeared on CIO.com