Skip navigation EPAM

Welcoming the New Age of Automation Assessment

Harsha Srikantan

Manager, Intelligent Automation Consulting, EPAM
Blog

Today, few technology solutions garner more attention and excitement in the business transformation space than intelligent automation (IA). This is largely due to its growth potential and ability to create simpler, more seamless customer experiences, as well as increase efficiencies and lower operational costs. Even though most implementations work remarkably well when the process being automated is relatively straightforward or backed by strong documentation (generally through robotic process automation or RPA), it is not uncommon for users to find an array of post-deployment issues. The source of these issues could be traced back to process analysis, where the mapping is incomplete due to one or many of the following challenges:

  • Unclear existing documentation — process maps or standard operating procedures (SOPs) of processes are incomplete, or important information is missing or out-of-date
  • Subjectivity of process steps — different subject matter experts (SMEs) may be associated with the process and share different interpretations of its inner workings
  • Increased complexity — process steps spanning multiple business functions, systems and applications are complicated
  • Outliers — edge cases or variations are often overlooked or unintentionally missed

These issues often result in failed automation implementations, which quickly lead to disillusionment with the technology in play or de-motivate employees in IT and business groups. Addressing this challenge of properly identifying processes prime for automation can give companies a better chance at transformation success. 

How to Overcome Challenges with Discovery & Selection

When you first start the journey towards IA, the basic question to always consider is ‘Should this process even be automated?’ Many processes or ideas that initially appear as great candidates for automation quickly fall short in terms of feasibility, returns on cost and effort, or alignment with the enterprise strategy. In analyzing these factors a little more closely, it’s important to dig deeper into the process and establish potential and strategic buy-in beforehand.

Today, most organizations focus on discovery and process selection methods based on mapping existing business processes through information procured from process SMEs, SOPs, and time or motion studies. In these scenarios, key information is often overlooked or missed in translation and additional steps or process deviations are not defined. SME subjectivity of process steps, as well as siloed information, leads to incorrect assessments about effort estimates or process run-times.

To mitigate these challenges, automated tools can be utilized during the discovery process to provide a more detailed analysis of the process. While similar to the concept of data mining, process mining focuses on extracting insight from process flows through existing transactional data. Process mining primarily utilizes event log information from large enterprise systems, like the CRM or ERP, and effectively analyzes large sets of operational data providing detailed, data-driven analysis related to key processes and their performance.

Event data typically holds basic details like the activity performed, when it was performed and how it was done. Using this audit-level data from actual transactions performed and recorded within enterprise systems, a process mining tool generates visualizations of the process with a step-by-step analysis of each individual transaction or case. Detailed views of the steps provide insight into three main areas:

  • Standard paths and the number of cases traversing this route
  • Bottlenecks where the flow takes more time to complete
  • Unexpected activities that are triggered for exception cases or deviations

The best advantage of using process mining is the focus it brings to issues within the flow or process. This view is a great starting point to re-imagine the process or re-engineer certain parts. Process mining also helps to identify and rationalize non-core and non-value-add activities while reducing manual reporting efforts.

Why is this Important Today?

Without elements of redesign or reimagination, processes will remain broken or inefficient. Bottlenecks and unintended loops need to be identified and addressed, or in almost all cases, automation will only amplify the effects of a bad process.

With advances in machine learning (ML) integrations, process mining tools can now be used not only for analyzing historical transactions, but even for monitoring procedures that happen in real-time. Consider a case where the system alerts the procurement team of a delayed supplier payment, triggering a cascading delay within the process. This should prompt a quicker resolution of issues within the process to mitigate delays. These cases have significant potential in the areas of anomaly detection and alerts. With the visualization features, business functions can also see how changes made to parts of the process would affect the entire process flow.

Process discovery tools are relatively new entrants to the IA space and offer the next iteration within the discovery landscape. Instead of relying on just process maps or process mining, process discovery tools capture actions that the user takes to run through the process and can automatically create process flows that detail the steps with the underlying applications. This shows great promise for rapid analysis of current processes and system interactions. Based on time and effort studies, automation potential and cost benefit can be easily calculated.

Conclusion

Process mining offers significant advantages in evaluating business processes and helps to identify inefficiencies of end-to-end processes, paving the way for strategic changes. With more transparency and fact-based visualization of process data, enterprises can trust that they are investing in the right areas prime for automation. As IA moves from more traditional use cases for RPA to exploring ML and cognitive solutions, data-driven decisions into automation investment and transformation strategy will lay the groundwork for eventual and sustainable enterprise success.

Related Content

Intelligent Automation

LEARN MORE


The Human Experience Component in Intelligent Automation

LEARN MORE


How Intelligent Automation Can Optimize Enrollment Season for Payors

LEARN MORE

Hello. How Can We Help You?

Get in touch with us. We'd love to hear from you.


Our Offices