How do you evaluate smart ERP systems?
21 Dec 20
Today the media is flooded with messages referring to concepts of Machine Learning, AI, Robotics, digital assistants and much more included in “Emerging Technologies”. And not least the vendors of ERP systems are hitting the drums that their systems are now becoming smart and self-learning.
And we can state that all of the global ERP vendors are investing massively in technology in order to be able to take the step towards a “smart” system. By this meaning a system that over time will be capable of understanding the customer’s business and thus take over parts of the user’s daily work. All this seems exciting, but the question is how far have the vendors actually come, in reality, and what can the already offer today in form of smart services?
The concept of digital assistants can be summed up as a combination of 5-10 different technologies that need to co-operate together for the user to experience the system as a “real” colleague with the capability of taking over parts of the daily work. We usually include speech and language interpretation, RPA, machine learning and predictive analysis as basic components that together with Big data, Data management and APIs create conditions to reach a level where the system (or platform) in collaboration with the outside world can be perceived as “self-thinking”. And most of these technologies are also mentioned when the global vendors are listing their prioritized investments for the future.
However, when it comes to actually presenting a list of all smart services that can be delivered “out-of-the-box”, the list from the ERP vendors often becomes very short. And in many cases embarrassingly short. Several ERP vendors even consider the today very basic function of “suppliers invoice matching” to be a smart service. But, presenting a system with capabilities of supporting smart purchasing, re-planning of logistics or production, based on data and patterns that no one else has – can rarely be found.
However, this should not be interpreted as criticism but as basic facts. In reality, most ordinary ERP vendors and systems have not yet left the starting pits. On the other hand, we see that the vendors of ERP systems are building frameworks outside or around the system and where new technology is built into the platform to create smart services. And with the ambition to, over time, gradually increase the number of services that can continuously process the ERP system’s data and put this in context with other external data to make predictive analyzes and propose measures to act on. SAP as well as Oracle, Infor and Microsoft have come rather long way in this area, and they intend to in the future be able to distribute this type of preconfigured services in connection with ongoing upgrades.
Is it then possible to define requirements of smart services when you plan to select a new ERP system?
Yes, it is possible to define requirements of this. But you must at the same time understand and be aware that the “smartness” more relates to a tool where the customer himself needs to work with the tool and trim it over time for the system to be smart. So, the requirements must be defined in the sense of a tool (or platform) and the possibilities for the user to handle this tool. And with focus of how flexible the tool is when it comes to combining data from several sources and not only the ERP system. To this must be added the importance of being an open system and database. We can state that even by 2020 we all too often encounter large and established ERP systems with major weaknesses when it comes to open or defined APIs to the outside world.
If you want to evaluate the ERP system based on “smartness”, you should list your 10-15 most important processes and ask the vendor to present, in live demo, how the ERP system will create automation of the customer’s business. Request a presentation of the tool used to define and build new business rules to control the processes and how the system can be configured so that it can learn to find patterns in daily work on its own and suggest changes.