How is artificial intelligence – and its prominent discipline, machine learning – helping deliver better business insights from big data? Let’s examine some ways – and peek at what’s next for AI and big data analysis.
How to identify Robotic Process Automation (RPA) opportunities
Robotic Process Automation (RPA) doesn’t suit every process. Here’s how to identify the best use cases in your organization.
How to construct a diagnostic framework for RPA
Particularly once you’ve generated a list of candidate processes for RPA, you can translate these kinds of questions into a more evaluative rubric.
“A diagnostic can be used to systematically evaluate candidate processes across all functions and verticals to determine which ones are fit for RPA,” Huff says. He recommends rating potential RPA fits on seven criteria in some way that makes quantifiable sense, such as low, medium, and high. (You could also use a numerical range such as 1-10.) Huff also suggests creating a heat map based on the results to help prioritize projects and create a roadmap, since you’re likely not going to tackle everything at once.
Here are the seven criteria:
- Transaction volume
- Prone to errors or rework
- Process predictability
- Rules-based exception handling
- Manual work involved
- System upgrade timing
- Controls importance
Four functions where RPA can deliver business value
Sometimes it’s helpful to see and hear accessible examples of where RPA has worked well in other organizations.
In terms of business functions or departments, Edwards from Eggplant notes that most any back-office function likely has processes that are good fits for RPA.
“It is great for automating back-office business tasks across industries, such as accounts payable, returns processing, and warehouse management, as it increases productivity and reduces errors,” Edwards says. Other examples include:
- Finance: Sales order processing, invoice processing, accounts payable
- Management reporting: Gathering information from lots of sources and putting them into consolidated XLS
- Marketing: list management, email, social, digital marketing, CRM management
- IT and/or HR: Staff onboarding and off-boarding
In terms of vertical industries, Huff mentioned several – such as banking, insurance, and supply chain – with abundant processes that fit the framework and questions above. Landreman points to the healthcare industry, which is Olive’s target market, as another.
“We’ve found that RPA can significantly reduce inefficiencies on the non-clinical side of our healthcare system – these technologies are perfect for many time-consuming, costly revenue cycle processes such as checking the status of claims, eligibility and benefit verification, and more. And it’s helping free up time for skilled workers to focus on tasks that require a human touch, instead of spending time manually processing data.”
Four categories of tasks that suit RPA
Finally, Bultman from Nintex shares several broad process or task categories that are department- and industry-agnostic, meaning they can apply to just about any organization or business function.
1. Updating information: “For example, organizations often clean up addresses in their system in order to get a reduced postage rate when sending out mailers. They export the addresses, pass them through an address cleanup service, then use RPA go into each account and update the address,” Bultman says. “Could this project be done programmatically? Probably. But that would require IT or a programmer to get involved. With RPA, the employee tasked with updating this information can use RPA to create an automated task for this themselves.”
2. Migrating information: “Organizations often need to move information from one system to another. For example, if a bank merges with another financial institution, the bank systems need to be consolidated,” Bultman explains. “Using an RPA approach, the data can be exported from the old system and keyed into the new system at superhuman speed. This approach is advantageous in that it does not bypass the rules of the destination business system – that is to say, the data being put in is formatted and stored properly every time. Other code-heavy approaches bypass the screens of the application, which prevents the work from being captured in the system’s audit logs.”
3. Urgent tasks: Any process with a trigger that requires immediate attention can be a good fit for RPA, which can often run that process faster than a human can. “If a bank is notified of a data breach that compromised a list of debit or credit cards, that financial institution wants to close those cards as quickly as possible to mitigate their losses and protect customers from the hassle of having their funds stolen,” Bultman says.
4. System monitoring: “RPA bots can be used to monitor critical business systems. If an online retail store goes down, it could cost a company millions of dollars,” Bultman says. “Organizations are using bots to simulate human interacts with their online apps to ensure things are running smoothly. If anything goes wrong, the bot can immediately report the problem in a variety of ways, including sending email, text message or Slack notifications, to name a few.”
Again, the possibilities for RPA may seem limitless, but asking the right questions and applying some type of evaluative or diagnostic framework can help you develop a sound strategy. Also, don’t confuse those seemingly limitless possibilities with “everything.” There are areas where RPA isn’t the best approach. Look for more coverage on that topic in a follow-up post.
[ Want lessons learned from CIOs applying AI? Get the new HBR Analytic Services report, An Executive’s Guide to Real-World AI. ]