Robotic process automation, or RPA, is the use of software with artificial intelligence and machine learning capabilities to handle high-volume, repeatable tasks that previously required humans to perform. Simply put, RPA technology enables companies to automate manual processes across the enterprise using software robots, in turn reducing human error and saving money on labor. One key area where this sort of automation will make particular sense in the near term is configuring and managing disparate components of an IoT deployment.
R20;Repetitive, manual work can be in many areas of any organization, and, mostly, it’s a strategic or commercial decision which areas in companies are being looked at first to eliminate the manual work,” said Devin Gharibian-Saki, chief solution officer at Redwood Software, a vendor of RPA software based in Houten, Netherlands. R20;That’s typically driven by whatever higher-level initiative there is. For instance, if there is an initiative to increase headcount in another area, that is typically a trigger for an RPA project.”
Saving money is a key driver for implementing RPA, but it’s not the only one, according to Gharibian-Saki. Many companies look at securing the data quality of information they are storing and information they collect, while others want to increase customer satisfaction or are looking to implement RPA as part of a general standardization initiative, he said.
“There are multiple higher objectives that can conclude in an RPA project,” Gharibian-Saki said. “My observation is that there is always a strategic decision first, and from that strategic decision there are RPA initiatives — some of which are also done on a very tactical level, i.e., there is a small team that wants to do something more efficiently and they just do a small project in their area.”
But the main objective is always to remove repetitive, manual work, he said.
Sarah Burnettvice president of research, Everest Group
Echoing this, Sarah Burnett, vice president of research at Everest Group, a consulting and research firm based in Dallas, said robotic process automation technology allows you to connect to systems through their user interface. That’s why it’s called robotic — it mimics what the human would do at the keyboard.
“RPA connects the underlying system through the user interface, and that’s the thing that’s made it so popular so quickly — you don’t need huge system integration projects,” Burnett said.
RPA technology can be used by any business as it is a horizontal application. It’s really more about what kind of processes RPA is geared toward and which are those that involve a lot of data transactions, Burnett said.
“There are telecom companies that are using it to do a lot of heavy, data-based transactional processes, for example; it’s ideal for those,” she said. “And we’re hearing lots of amazing stories about returns with RPA, so upward of 30% efficiency savings because the robots do the jobs of two or three people.”
Robots can do repetitive rules-based tasks quickly, so they process transactions faster. Because the robots are rules-based, they also reduce error rates as they do exactly what the rules tell them. Taking that a step further, companies can also embed regulatory compliance in the rules so that every step a robot takes generates a robot trail and everything is recorded for posterity.
Challenges of robotic process automation
Despite the benefits, there are some challenges around implementing RPA technology, including finding the right projects to automate.
“Because of the massive hype about RPA, organizations think they can take it on and deploy it very quickly,” Burnett said. “But then they find there’s a bit of learning and it doesn’t go as quickly as they thought it would.”
Despite this, she added, companies are getting returns on their investments within three to six months, which is much faster than with a lot of other software.
Another challenge organizations face after implementing RPA is figuring out how to scale up. “Companies deploy a handful of robots and say, ‘Great, this works, but what do I do now? How do I scale up to the rest of the organization?’ That’s because it’s an immature market and we’re still learning about robots and how they work,” Burnett said.
RPA and IoT should become best buddies
With the increased data and information-sharing brought about by the internet of things, organizations have been able to better streamline business processes and make operations more efficient — making IoT a prime candidate to pair with RPA.
In particular, RPA can be critical to automating organizations of all shapes and sizes, providing the platform from which to manage IoT devices, a task that might otherwise have to be done by employees working in the back office.
Together, RPA and IoT can enable companies to optimize management processes as well as employees. Offering organizations the means to access new business intelligence as well as the ability to manage it, these technologies together can allow companies to improve operations by, for example, providing better customer service or better managing how products travel through the supply chain.
In one application, said Burnett, there’s an airline using IoT data in its luggage handling process and then using RPA technology to bring that information to the operators who can decide how much more cargo they can take on board the flight.
“RPA is a disruptive trend in digital transformation because it’s fairly fast and easy to implement, has almost immediate outcomes and payback periods are often six months,” said Bobby Patrick, chief marketing officer at New York City-based RPA vendor UiPath.
The first wave of RPA and IoT over the last 12 to 18 months, Patrick said, has been in the back office around finance and HR operations, accounting operations and procurement.
HP Inc. recently created a robotics and automation capability and strategy for finance using UiPath’s RPA technology.
“We’re deploying RPA in the finance space, specifically controllership,” said Michael “Chet” Chambers, senior director of innovation at HP. “We’re looking at cash outs, and we’re looking at posting invoices and several different areas on a global scale. And when we automate functions in one region, we’re able to extend that automation, with some minor changes.”
One clear benefit, Chambers said, is that using UiPath’s technology, his department created robots that can open email containing invoices.
“We’ve taught the robots to read the email, extract the correct information from the invoices, bring that information into the finance system, do a three-way match — the purchase order, the invoice and the work order have to match — and post it,” he said. “Some bots are set up to read the invoices in the system where the humans would have to manually enter the invoices. Now, there’s no human intervention involved in paying an invoice.”
RPA technology, Chambers said, enables the company to free up employee resources to do more valuable work.
“There is also a pickup in the use of RPA in the front office now, whether in contact centers and customer service areas or in areas around supply chain and manufacturing,” Patrick said. “It’s the latter where we see IoT today.”
One of the reasons robots are so important to IoT, he said, is that most big companies have legacy systems and even operate on spreadsheets when it comes to processes that need to support all the data coming from sensors on equipment. “Robots can speed the ability to adopt new technology at the edge while being able to integrate that technology with existing systems and processes. So, that’s what we see today in oil and gas, and in postal operations around the world as well. And 2018 will be a year where we see much more around IoT integration and applying AI to those processes as well.”
In the future, HP plans to use RPA technology to capture data from the images of thousands of paper invoices that are physically scanned into printers at one of the company’s three major scanning hubs across the world every day.
“We capture those images now and we’re training [optical character recognition] software to work with our bot to be able post those to make them actionable data instead of pictures,” Chambers said. “I also have a drone program that I’m working on to do physical inventory. The bots are going to read the data from the drones, extrapolate what it is a picture of and what pieces of it are pertinent, and post it to the correct area.”
DIY RPA software?
Robotic process automation technology as provided by Automation Anywhere Inc. enables organizations to create their own software robots to automate any business process, said Abhijit Kakhandiki, senior vice president of products and engineering at the company based in San Jose, Calif.
Automation Anywhere offers cognitive bots that learn as they work, as well as an analytics platform that enables businesses to aggregate and manipulate the data handled by the robots in a centralized analytics dashboard.
“For example, JPMorgan Chase, which is one of our banking customers, has one bot that actually talks to 61 systems, which is a lot of systems for one bot to talk to,” Kakhandiki said. “What that means is that in the context of one business process, all these systems are coming together to get the work done. That’s the difference between RPA and just your system-to-system API integration.”
As the use of RPA software grows, businesses want it to play nice with their enterprise software as well as with all the ways in which they are getting data, he said.
“One thing that differentiates us from some of the other vendors out there is that we also have our business intelligence platform that is on top of RPA,” Kakhandiki said. “The bots take in and log data, and through this data, you can now surface new business growth insights/problems to the humans.”
In this context, IoT can act as a potent data source, Kakhandiki added.
“IoT can be a great trigger for triggering a bot to do something,” he said. “So, all the different IoT platforms that are out there can use our APIs to trigger that bot. And that bot can go into any complex system and be able to enter data into a system or modify data — all those kinds of things.”