Can you imagine driving your car blindfolded? You may even succeed with good guesswork and a solid exposure of surroundings! But is it really reliable, and more importantly is it worth trying?
Business processes are a similar case. It needs special acumen to see what exactly is happening inside, where are the bottlenecks, should it get automated, what needs to be done to maximize efficiency and amplify value. It must not be based just on guesswork or SME’s biased view! Organizations, for decades have been striving hard to gain insights and transparency to understand the as-is processes to identify gridlocks in the flow and create a baseline for improvement.
Digital advancements across industries also assuage enterprises’ efforts to model virtual view of the processes. There are products which consume logs generated by the disparate systems, capture actions performed by individual users, and analyze them using AI algorithms to create as-is process maps. Crafted on the foundation of empirical data, these process models paint the end-to-end process flow comprising of multiple sub-processes and tasks executed by various actors with associated lead times.
The depth and granularity of these process models vary based on the business need, and hence, the way to capture such information will vary accordingly. It ushers in a brand new opportunity of process discovery which in contrast to process mining, captures more granular control level actions at each step of the task.
Process mining products heavily rely on analyzing event logs captured from enterprise systems. It needs to build a data warehouse, integrating the logs from these applications. Process discovery products record users’ click-stream across systems and with the help of AI, they reconstruct the virtual map of what users do. Since it captures human actions on screen, process discovery products may miss seizing the non-systemic data (like decisions based on visual inputs) overlooking potentially relevant steps.
Process mining, for example will give an overview of a procure-to-pay process showing the execution time of individual tasks such as purchase requisition, vendor selection, purchase order creation, invoice reconciliation etc., the lead time between these tasks, and different ways in which users are traversing to execute this end-to-end process.
On the other hand, process discovery will help create business requirement documents for a robotic process automation project to automate a particular task (say purchase requisition) providing as many details as possible focusing on:
The RPA tools available in the market today operate at automating step level activities of individual tasks. Most often, a single actor executes all the steps of these tasks.
Balancing the right level of granularity of process knowledge is crucial to amplify the process value. And if you want to accelerate your automation journey, automated process discovery will certainly be your first step.
So, would you prefer driving the car blindfolded or enjoy an intelligent driverless automatic car instead?
Just sit back and relax!