Most of the projects we do here at Kien have to do with enabling our client businesses to work smarter, faster and excellent. Therefore, business process automation plays a very important role in almost everything we do. We use business process management systems (BPMS) such as Bizagi as one of the core enablers to build what we call speed companies, but in practice, throughout the evolution of these kind of speed projects, we have to deal with many different challenges, which cannot be solved by “vanilla” BPMS.
A real-life use case
Imagine a B2B use case where the client (let’s say, a department of a large engineering company, responsible for warehousing, asset maintenance and logistics support) wants to improve the way (internal) customers send material reservations to them and how the fulfilment is done.
Currently, there is a team dedicated for order entry, which come in mainly through e-mail and telephone. The actual order entry is done in the ERP backend, and from there on, the process of fulfilment goes forward semi-‘automatically’.
However, there is little transparency across the chain for both the customers and the internal team and the level of process control and quality management is very low. In the past, some proof of concepts were done to redesign the business process, but the integration with the ERP system repeatedly came out as the key bottleneck, rendering much uncertainties and continuity risks to the organisation, resulting in low trust to push these projects further. Then BPMS comes around the corner.
Igniting digital transformation
A business process like this serves as good example for candidate flows (so called ‘seed processes’) for BPMS to get track as a rapid digital transformation tool. Imagine that what you design as a process model (for example, using BPMN as the modeling standard) can be transformed into real working applications and vice versa, whenever you want to change the process model.
As long as the candidate process creates a business impact, meaning efficiency improvement (more speed and less costs) and/or customer satisfaction (or anything else that is part of Lean thinking) the “oil stain” effect could take place, being the concept of process automation spreading across the organisation, connecting the silo’s, improving quality and after a while, creating more headspace and resources for innovation.
Now, this is the theory: in larger companies which have bit of history, reality is much different. Let’s have a look at the full eco-system that BPMS needs to connect with in these cases, as to make a real impact.
The challenges when implementing BPMS
First of all, in general, BPMS systems do not primarily focus on user experience (UX) as a primary goal/objective. They help you to automate your business flows. That means that it might not be a very good idea to open up your BPMS portal to your direct customers, however, you need them as they are the key actors of your business process.
What often happens as a result is that the underlying BPMS logic is embedded either in an existing toolset or platform (like an intranet portal page) or that specific applications are built for that purpose.
In our example, a dedicated B2B application was created, basically serving as the customer-facing product catalogue and ordering system, using the APIs from the BPMS to send orders and get order statuses. This kind of solution, although very logical and handy, adds an extra layer of complexity.
From there on, all goes well within the scope of the BPMS. Using the case management concept, the process flows forward and back, taking business rules into account, serving the organisation with workflows, sensors, analytics, and much more automations and makes the process flows run super-fast, during the business process runtime.
However, once data has to be persisted you run in to problems. Although certainly possible, the goal of the BPMS could never be to hold master or transactional data itself (respecting the key principle of “separation of concerns” ensuring agility of the IS/IT landscape) – therefore you want to store the data in the source systems. In our case, an ERP system.
Integration is the key bottleneck
So there you go, you end up with the same set of problems as mentioned earlier: integration is difficult, expensive and unfortunately, the key inhibitor of change. This is not really due to the BPMS itself, but because of the fact that you need to build new integration capabilities between the BPMS and the source system(s).
And yes, of course many companies have middleware layers or even full blown enterprise service buses in place – but, really, have you ever seen them being used without major hick-ups and high costs involved? That’s exactly what you wanted to avoid with the introduction of BPMS.
And it gets even worse when you need to connect with legacy systems that lack the basic integration capability or are just so outdated that it is very unwise to add new features to it. These kind of situations will instantly kill your BPMS opportunity and the problem is – you can’t really explain it, as your key audience most probably will be “business guys” who may lack the technical skills to understand the problem (“How hard can this be?”) and whose opinion might be more tenacious due to past experiences dealing with these kind of problems. Your project will be killed without the chance for a next one.
Robotic Process Automation and its magic
Now think of the following. Imagine you do not have to build any technical integration, but that you are able to use the same screens that are currently used for manual order entry in to the ERP system. You add a virtual worker to your team.
Instead of “replacing human work with robots” (which more or less is the key proposition currently used for using RPA and, in my opinion, a very negative one), also called “Digital Labour”, you amplify your solution to overcome the legacy bottleneck, enabling your teams to focus on really important stuff. No need to work through technical integration scenarios or complex data mappings: script the scenario in to the software robot and connect it to your BPMS API.
Basically, this is how it could work:
The key objections to this concept either are in the category of “This is a silly work-around, you are not fixing the real problem” or of the likes of “Screen-scraping has been out there for a very long time and has not proved itself really in end to end business scenarios, so why would this work?”.
However, you actually áre solving the key problem, which is IT not being able to deliver value fast, or legacy technologies not being able to follow the speed of change, resulting in to endless and expensive integration projects.
For the screen-scraping argument, RPA really is something else. With modern RPA-tools, you can build very complex scenarios, including intelligent feedback loops with data from the source systems and advanced orchestration, just as you do with the BPMS part of your solution. In fact, RPA can be seen as “screen level”-process management, serving very well as a powerful extension of a business process automation solution.
At KIEN we use Bizagi BPMS in combination with Automagica or UIPath to realize these use cases.