In the modern enterprise, there are many processes that are ripe for automation—with hundreds of processes from which to choose, the challenge for most organizations is identifying the ones for which automation will deliver the greatest benefit. As organizations embrace robotic process automation (RPA) to meet heightened customer expectations, work more efficiently and create a better employee experience, process discovery is playing an increasing role in delivering on the promise of RPA.
Simply put, process discovery is how organizations discover how any process in the company is executed. They can then identify and prioritize the processes to improve or, in many cases, to automate. Some types of processes—for example, ones that are repeatable, definable and rule-based—are more suited to automation than others.
Traditionally, organizations have identified processes for automation by brainstorming with business stakeholders, observing how employees do their jobs and interviewing experts or asking them to record themselves executing a process. These approaches aren’t without their drawbacks, though: Because they rely on human judgment and trial and error, they are subjective, expensive and time-consuming, and they fail to capture how one employee performs a process differently than another.
Today, however, a growing number of organizations are automating process discovery itself and removing many of these roadblocks. They’re leveraging machine-learning-based tools to identify key business processes, record variations and prioritize processes for automation in addition to designing automation workflows. By taking an objective, data-driven approach to process discovery, organizations can reduce costs and risks while boosting performance and quality. Automated process discovery also enables organizations to deploy process automation more quickly and helps ensure that no prime opportunities for automation are overlooked.
The Role of Process Discovery in RPA
”’Automation’ is an old term that has been used to describe virtually every type of computer system, from general ledgers to manufacturing robots,” Thomas H. Davenport wrote in Harvard Business Review. “The focus of automation tools was at one point to automate structured, predictable workflows typically within a specific domain, for instance in IT or marketing. But one automation tool, so-called robotic process automation (RPA), has become a generalized tool for executing structured workflows, particularly for processes that involve data from multiple information systems.”
The use of RPA is growing by leaps and bounds; Fortune Business Insights estimates that the global RPA market size will reach $7.64 billion in 2028, up from $1.61 billion in 2021. Growth is being driven not only by the effects of the COVID-19 pandemic, which increased pressure on businesses to automate back-end and front-end processes, but also by advancements in artificial intelligence and machine learning. The BFSI industry leads in adoption, with increasing use due to regulatory reporting requirements and balance sheet reconciliation.
“The key driver for RPA projects is their ability to improve process quality, speed and productivity, each of which is increasingly important as organizations try to meet the demands of cost reduction during COVID-19,” said Fabrizio Biscotti, research vice president at Gartner.
RPA costs one-third as much as an offshore employee and one-fifth the cost of on-site staff, and can cut costs by 25-50%.More than four in 10 organizations are already using RPA; in the contact center, RPA is enabling customer service leaders to deliver better customer service by standardizing and accelerating agent work to improve the customer experience; integrating applications to reduce errors and improve compliance; and increasing agent confidence so they can better nurture customers, according to Forrester.
Regardless of industry, the success of an RPA deployment hinges on being able to quickly and accurately identify the right processes to automate; the Shared Services and Outsourcing Network (SSON) has estimated that nearly half the automation projects that fail do so because of wrong choice of processes.
The Benefits of Process Discovery
Process discovery enables organizations to:
- Cut costs by reducing variations and the errors that lead to unnecessary costs and rework.
- Pinpoint the processes to automate to maximize the bottom line.
- Eliminate the guesswork in identifying a process to automate, mapping how it is done and designing the automation workflow. This improves quality and performance while removing human bias.
- Find process sequences or variations that are not obvious to the human eye.
- Increase visibility of ownership for specific steps in the process.
- Maximize the ROI of robotic process automation, including the ability to:
- Increase throughout (robots are four to five times faster than humans and can work around the clock).
- Ensure compliance (robots can work with 100% accuracy, and RPA can help companies uncover non-compliant interactions and take action more quickly).
- Increase employee engagement (by freeing employees to work on higher-value tasks rather than repetitive work).
- Reduce costs (robots can do the work of 100 people).
- Scale as needed (not only in terms of the type of task to be automated but also the number of tasks required for any given goal).
Challenges Associated with Process Discovery
On paper, process discovery is a simple set of steps, but roadblocks commonly emerge in practice. Challenges associated with process discovery include:
- Identifying the right level of detail, including understanding when rare occurrences are important and when they can be overlooked.
- Utilizing incomplete data, which can prevent the organization from uncovering process variations.
- Relying on historical data, which means process discovery can fail to identify things that could potentially take place.
- Getting the right people involved, including stakeholders from across the organization who are involved in the process.
How Process Discovery Differs from Process Mining
Process discovery is a subset of process mining, which is designed to help organizations identify bottlenecks, gaps and inefficiencies and optimize their processes. Other areas encompassed by process mining include conformance checking (monitoring deviations by comparing model and log), social network or organizational mining, automated construction of simulation models, case prediction and history-based recommendations, according to the IEEE Task Force on Process Mining.
Process mining uses specialized data mining algorithms to identify trends, patterns and details contained in event logs recorded by an information system and automatically turns them into objective visualizations of an organization’s processes. In contrast to process discovery, it reads data from system events, not user-level tasks; most process mining software, however, includes process discovery capabilities.
Process mining enables organizations to do three main things:
- Discover process models (e.g., create a new, optimized map of what a process should look like).
- Compare an actual process to an ideal process.
- Better understand an existing model with new information (i.e., performance or cost).
While process mining studies the processes that are in place, process discovery studies how humans execute those processes by recording user interactions with systems, analyzing repetitive actions and creating bots to automate those actions. Process mining is also subject to some key limitations, including that it only works for systems that produce logs (so, for example, not for Excel, Microsoft Teams chats and many other productivity tools).
Process discovery and process mining both use automation technology to identify and map business processes, but they differ in several key ways:
- Bias: Process discovery is based on empirical data; process mining is subject to human bias.
- Compatibility: Process discovery is compatible with most software applications; process mining is not compatible with all software applications.
- Data: Process discovery captures task-level data; process mining captures system-level data.
- Integration: Process discovery enables easy automation from mapped data; process mining requires additional integration with RPA.
- Monitoring: Process discovery continuously monitors processes to ensure real-time analysis and enable rapid retraining in case of changes in the process; process mining does not.
- Timeline: Process discovery can be completed in a few weeks; process mining can take several months.
- Workflows: Process discovery enables functional automation workflows; process mining does not.
How to Choose Between Process Discovery and Process Mining
Process mining tools are largely used by organizations that want to optimize their process, conform existing processes to certain specifications, create harmony between distinct processes, or get future predictions regarding their processes. AI-based process discovery tools, on the other hand, are largely used to identify RPA opportunities for automation, enable digital transformation, unveil previously unknown processes for in-depth process mapping or understand a process (without integrating anything).
“Both process mining and process discovery play a crucial role in assisting a true digital transformation for any enterprise,” Sonika Aggarwal wrote in TWDI’s Upside magazine. “One unearths the way your system applications work; the other focuses more on the human aspect and how your employees interact with the systems to unleash the dark and invisible processes.”
In deciding whether process discovery or process mining will better meet your organization’s needs, ask yourself:
- Do you want to integrate the results with your RPA initiatives? Process discovery enables easy automation, while process mining requires additional integration with RPA.
- Do you need to identify invisible processes as well as the visible ones? Process discovery sheds light on the things users do that diverge from established processes.
- Are you looking for actionable insights? Process discovery not only captures how users perform processes but also provides the information needed to implement changes that will have the greatest impact on the organization.
- Do you want to build an ideal business process model that can make existing processes more efficient? Process mining may be the best fit.
Shorten the Path to RPA
RPA is fast moving from a competitive differentiator to table stakes in many organizations, and process discovery helps shorten the path to automation. In the contact center, process discovery is already being used to record customer interactions, enable real-time access to customer information and much more.
An automated, data-driven solution – such as NICE’s Automation Finder – is the key to unlocking RPA ROI. By automating process discovery, organizations can leverage intelligent cognitive technologies such as desktop analytics and unsupervised machine learning to analyze employee desktop data, including keystrokes, mouse selections, applications used, pages visited, field entries and handle time. It enables businesses to continuously map and prioritize processes for automation, ensuring a reliable stream of RPA friendly-process opportunities to consider in their pursuit of ever-greater operational efficiency. Read our RPA Guide or learn more about how NICE is bringing people and robots together with automation.
