Five Questions with Martin Szenig
VP of Technology and Process Optimization Studio Partner at Globant
Who is Martin?
Martin started doing manual consulting work at different companies over 15 years ago and kept thinking that there must be a better and more efficient way to do this work. Ultimately, this led to him implementing automation and lean process improvement as he wanted to increase the value of his actions. Today, he takes a hyper-automation approach to consulting for companies, which includes intelligent automation process mining, BPM, BPMS, and associated toolsets.
Is there a standard approach for Process Mining?
When Globant engages with a new client, they have a standard approach, but at the same time, every client is different and there is no “silver bullet” for how to do these kinds of actions in the “best” way however, in general, the following steps are used:
● Educate: First, we demonstrate the potential and explain what process mining can do. Sometimes people may not understand why this technology is relevant and what we can do with this information. We don’t want to implement just to introduce new technology; instead, we want to ensure process mining is adding value by doing it the right way.
● Identify Use Case: Next, we need to find a great candidate where we have the information, people to support the project, and a gut feeling that there might be valuable outputs which can drive business transformation.
● Find Data: After that, we really need to get more tactical as we look to outline what we are going to map and what data sources we are going to go after.
● Industry Considerations: Depending on the industry, there are some differences in where there is the majority of value. Most companies see value in “Order to Cash” and “Procure to Pay,” but then you get into use cases specific to the industry. A good example of this would be Meter to Cash processes in the energy industry where there might be less pre-made analysis and more custom-driven insights.
What is One Example of a Unique Use Case?
At Globant, they have done a lot of “standard” process mapping use cases, such as SAP, JDE or WMS systems, but beyond that, there are a lot of different custom use cases that can be applied.
Globant has used process mapping on a custom software which combines SAP, Excel, and IoT devices to provide a full end-to-end process view. This was a very unique data model, as they first needed to build a common database with all these disparate data sources. What is cool about this use case is it is tracking a person’s journey when visiting a sports stadium. It allows them to understand when a customer bought the tickets, when they checked in, and then combine that data with technology like face detection to understand their experience at the stadium and improve their experience in real time.
Where Do You See Technology Evolving?
Process mining is part of a broader suite of tools supported, which includes RPA, Machine Learning, and BPM. All of these technologies have different use cases and sweet spots to add value. Over the past few years, we have started to see these technologies converge as platforms are starting to support all these different capabilities. The problem today is the use cases typically do not easily leverage the best of these different pieces. There might be technical hurdles to bring in these different capabilities, cross-team conflict for these broader use cases, or data limitations. As a result, today Globant is leveraging separate models in different tools, but in the next five years I think we will see more automatically triggered modules crossing between process mining, RPA, and AI/ML. In essence, we will have a single autonomous model able to leverage the best of these different technologies, which will enable real-time actions and improvements.
What Are The Main Process Mining Use Cases You See?
There are four main use cases that clients normally use when it comes to process mining:
System Migration - Leveraging process mining to prepare the current state process maps and then using this information to build the new system and prepare for migration.
Operational Efficiency - Most popular use case in understanding how stand-alone processes are performing and how those processes are interacting with other processes and systems.
RPA Identifications - Determining where we can best automate our existing processes, track improvement gains, and ensure our automations are performing as expected in the context of the end-to-end process.
Audit & Compliance - This is one of the fastest-growing use cases as we work to understand where compliance is failing, the specific path of those orders, and uncover potential fraud. Using AI with process mining to do more proactive and reactive analytics on the processes to help enable efficient audits.
Besides these four use cases, there are a lot of specific use cases that are customer- specific, such as tracking waste or CO2 emissions. Over time, we expect these company or industry-specific use cases to become more standard and popular as more companies look to track different parts of their operations.
How Do You Find Value?
Process mining is a great tool to monitor processes as you are able to find an opportunity, implement a change, and then measure the improvement afterwards. The challenge, however, is that principle process mining by itself doesn't execute anything. You are only finding the opportunity, so operating alone will not have an ROI. However, sometimes you could not find that opportunity without process mining, and even if you could, how would you track the impact without process mining?
This is very different from something like RPA, which is very straightforward because you reduce the time, you reduce the cost, and all of a sudden you’ve got a clear ROI which you can communicate back to leadership on the project’s success. However, with process mining you get that initial insight, then you can layer on RPA improvements and ultimately then bring in AI to start improving the process proactively. This is a much more complex program potentially involving multiple technologies and teams, but has a much higher reward than a simple automation.
A good example of this would be looking at shipping. You could start applying process mining to a warehouse system to understand inefficiencies in the process and identify potential areas of automation. You could apply RPA to some of the simpler administrative tasks to further enable the shipping process. Then you could think about how to embed AI/ML into the business process and do things like trigger automatic RPA workflow when an order could be potentially late.
I typically do not talk about a specific ROI that process mining can achieve, because this gets back to the individual use cases and, as mentioned previously, process mining is at least initially uncovering how work is being done. In general, we can get some insights in 4 to 6 weeks, but then it might take longer for the RPA and AI/ML work to be completed.