In Part 1 of this article, we had identified that Human bias is a major cause of failure in enterprise automation implementation. We had come to the conclusion that an automated, objective and data driven execution will definitely lead to successful automation.
The solution to effective automation change management
Currently automation CoEs are formed to enable different units of an enterprise to align with the overall business automation goals and ensure cooperation between them. The CoEs alleviate some of the cross unit communication issues that arise in a large enterprise, but do not address the process level gaps that tend to derail most projects.
As discussed above, the solution to overcome the risk factors in a transformative RPA implementation should be data driven, objective and automated. Also, it is important to understand that organizational improvement is an iterative process.
Thus, a three-step plan is necessary for effective automation change management. These reduce the risk involved in planning major changes in an organization and ensure the best Return On Investment (ROI) on automation implementation. The steps are:
Empirical mapping of processes
“If I had only one hour to save the world, I would spend fifty-five minutes defining the problem, and only five minutes finding the solution.” – Albert Einstein
With the complexities defined above, mapping a process could be an exhaustive task for consultants and SMEs. Even employees working on the process find it hard to effectively map their own work. With inherent human biases and inefficiencies, the task is arduous. And any holes in the mapping tend to snowball to cause more issues downstream.
Process mining has been the industry’s answer to this problem so far. Process Mining is done by algorithms that read process logs from the software used in the enterprise and build maps. Mining for this data has become more and more efficient over the years and is a significant step up from the manual process earlier. But Process Mining has its own disadvantages, namely how it handles multiple tasks running on different tools and the wait time for making some decisions.
This is where Process Discovery has its advantages. A Process Discovery tool can unobtrusively read keystrokes and record data on how employees accomplish their day-to-day tasks. The data thus collected, will be comprehensive and capture all variations and exceptions for the particular process. This data is free from human bias and covers multiple tasks, and thus provides automation experts and tools viable data for automation. This is what makes Process Discovery a game changer in the automation industry.
Analyzing and understanding the empirical process data
It goes without saying that a large amount of data needs to be collected and analyzed to build process maps. The analysis needs to be both thorough and fast to provide value to an enterprise. Advanced neural networks are required to process the multiple decision-making processes a human employee does automatically during a work day. The data needs to be further reduced to build process maps that are efficient, as the main goal of automation is increased efficiency over manual implementation.
As enterprises try to find what processes to automate, it is also important to know which processes shouldn’t be automated. There are instances where automating a process introduces additional complexities and delays and is therefore detrimental to the overall goal. The analysis done should therefore cover the following questions:
Implementing automation and further analysis
The steps defined above enable a smoother implementation of automation, but enterprises still need to be careful with the actual implementation. The automation tool chosen should not only cover the current scenario being automated, but also be scalable in the future. With the pace of technological innovation, Intelligent Automation should be the preferred mode of automation for enterprises.
It is important to understand that automation is an iterative process, and additional inefficiencies can be found and resolved down the line. If the automation tool is rigid and cannot scale or handle multiple bots, then there will be additional costs incurred to retool and reconfigure the whole implementation.
Process Discovery- the first step in your automation implementation
In summary, effective change management in the modern enterprise requires an unbiased and unobtrusive tool that can comprehensively map processes, analyze the data to build efficient process maps and also scale to handle additional processes in the future. Any enterprise looking to automate should therefore move beyond the traditional approaches like SMEs, consultants and Process Mining, and implement Process Discovery. Even enterprises which have invested already in automation tools need to rethink their process mapping to get the best out of their automation implementation.
Remember, change management starts with proper planning. And proper planning cannot be done without a thorough understanding of current inefficiencies. And the best way to identify inefficiencies, both known and unknown, is by Process Discovery.