In increasingly competitive markets, adapting is the only way businesses can stay ahead. In order to be able to adapt, organizations need to be able to move quickly and efficiently.
In a world marked by technological disruption, efficiency has taken center stage.
Until recently, business leaders only had a few, expensive options to handle efficiency problems; they hired more people, bought a third-party solution, and/or customized their enterprise resource planning (ERP) system. Robotic process automation (RPA) presents an opportunity to transform the processes that lead to inefficiency.
With RPA, businesses can use software robots to create shorter development cycles and increase robustness, for a smaller price tag.
Today’s leaders are learning how to achieve business outcomes by dividing work between people and robots, using their respective strengths accordingly. For instance, many processes involve repetitive, tedious tasks as well as higher-level, well-trained cognition. When organizations distribute tasks appropriately between robots and humans, the results can be staggering.
In this article, part one in a two-part series, we will focus on three unattended automation scenarios in which organizations run processes behind the scenes, with little to no human intervention. This video gives a quick reminder of how unattended and attended automations differ:
Part two in this series will focus on RPA use cases for attended automation in which employees work with robots collaboratively, in real-time on their local PCs.
Here are three unattended automation scenarios where robots work independently to initiate and complete processes.
The need to process massive amounts of data presents one of the clearest use cases for RPA. In this example, we’ll focus on some of our health insurance companies that process large volumes of claims and invoices.
Data is stored across multiple systems, some legacy and some modern, requiring employees to move information from system to system, updating and verifying along the way. Automation is difficult because many systems don’t have APIs that allow for easy integration.
With RPA, health insurance organizations, and companies with similar problems, can automate these processes to run without any need for human input.
Typically, one or more robots will sit on virtual machines or on back-office computers and run an unattended automation. Administrators can set the schedule in their RPA platform and allocate the optimal number of robots to process the claims and invoices and meet the business’s service-level agreement (SLA). This can be done easily in UiPath Orchestrator. Humans are uninvolved, except to handle rare exceptions that Orchestrator will surface to designated users.
Meanwhile, best practice coding principles provide complete, accurate audits for compliance.
With robots working continuously, humans are free to perform higher-level activities. Humans that previously spent their time moving and validating data can now focus on initiatives that require deeper thinking, reacting quicker to shifting demands from policyholders. Conversations with customers have the time and space to develop, and employees can focus on creating a better customer experience.
Some processes need human input to start, but require slow, manual labor to complete. With RPA, organizations can involve humans only when they’re necessary.
Humans can focus on high-level thinking, knowing that robots will guarantee delivery with best-in-class documenting and notifications.
Here’s an example of how a mortgage company uses this type of automation:
In the United States, the typical home-buying process requires lenders to routinely order mortgage appraisals that provide unbiased estimates of each house’s value. Despite their importance, ordering appraisals is repetitive and prone to mistakes.
Worse yet, if a mortgage appraisal is ordered late, it can delay the entire loan process and even risk incurring repercussions from the Consumer Financial Protection Bureau or other government bodies.
The problem of appraisals comes down to the inherent limitations of human labor.
Human logic is necessary to start the process, but completing it requires repetitive, manual steps. With RPA, a human can start the process by designating exactly when the appraisal should be ordered, specifying the appropriate add-ons, and choosing the appraisal company.
A mortgage appraisal might only take ten minutes to complete, but the real value for lenders is saving hundreds of thousands of dollars by ensuring fewer loans require escalation because of delayed appraisals. Meanwhile, employees can complete orders faster and work more closely with homebuyers, ensuring customers get the answers they need.
Dive deeper into the example (above): The Fully Automated Enterprise Is Here (and This Is What It Looks Like)
Many inefficient processes don’t present clear RPA use cases because they require intermittent, but necessary, human judgment.
Take an accounts payable (AP) department, for example. AP departments are often stuck repeatedly processing invoices and uploading selections of data to an ERP like SAP. The majority of invoices that AP departments process are non-standard, meaning they could be paper documents, faxes, or PDFs—filled with unstructured data in different formats.
The manual labor is burdensome and slow for humans, and the complexity makes processing invoices challenging for traditional automation tools.
Using optical character recognition (OCR) and machine learning (ML) models, RPA can interpret these invoices and pick out key information for processing. Occasionally, a robot may not be able to match a particular piece of data to a known data type (like invoice amount). Rather than fail the process and make a user review the entire invoice, the unattended robot can signal to a human that input is needed. In this a human-in-the-loop (sometimes referred to as "human in the middle") scenario, the human can then review and complete the tasks before letting the robot take back over.
With this use case in mind, the AP department can implement an automated, unattended robot with a human fail-safe. The unattended robot uses OCR to process invoices and upload data to SAP, but if the OCR can’t clearly see a field, the robot can notify users via email that the process needs their verification. AP employees only need to step in as necessary, so they can spend their time on more critical tasks, such as budgeting and planning.
As the human clarifies the exceptions mid-process, those changes are fed back into the ML model, which learns from the human’s activities. Once the model retrains, it won’t need human intervention to handle that scenario the next time it comes across it.
A process that previously seemed impenetrable to automation actually presents a compelling RPA use case. With a human in the middle, you can reduce human involvement to validation, and then ML can reduce it even further.
Taking full advantage of automation requires organizations to reframe how they think about their processes and the labor required to complete them. If you think 'automation first,' digital transformation can be an attainable goal rather than an aspiration.
The workforce of the future is going to be a hybrid of human and robot labor, and the most effective organizations will approach each problem by first considering the degree of automation they’ll use. Organizations excel thanks to their people, and the more that organizations can free their human workforce to do creative work, the more they’ll succeed. A hybrid, digital workforce enables humans to be more human, and more successful, than ever before.
Stay tuned for part two to learn how attended robots can work alongside humans to enable automation in even the most dynamic processes.
Want to learn more? Join us for our "Attended vs Unattended Automation" webinar