Unit4’s Self-Driving ERP™ in Practice
Posted by Robin Schaffer
Industry analyst, Mint Jutras, has published a report looking at the merits of Unit4’s self-driving ERP™. It says that it has the potential of being ground breaking. As well as providing a great overview of what we mean by self-driving ERP, it details some in-practice examples where services organizations can benefit every day from manual time and effort reductions, productivity gains and customer service improvements. One example focuses on a professional services firm and gives a fascinating, real world look at how this approach to enterprise computing will pay dividends.
“Projects are common in many services organizations. For some, projects are simply internal. But in many companies, particularly in professional services organizations, these projects are core to their business. Unit4 has been listening to these types of customers as they expressed a desire for better ways of winning profitable business. When your business is project based, that means coming up with more accurate estimates faster…
…To better understand this endeavor, put yourself in the shoes of a project manager at a project-based business that has identified a new opportunity. You need to come up with an estimate of cost, resources and schedule in order to propose a price that is both competitive and profitable. And you need to do so quickly and efficiently or either your window of opportunity will close, or your current projects will suffer, or both. If you are smart you don’t start completely from scratch. Instead you find a similar project, hopefully one that was successful, and start from there, modifying it to reflect the current needs of your prospect. Sounds simple, but in reality, how do you go about finding the right project to use as a starting point, especially if it was a project in which you had no personal involvement or experience? Unit4 is developing a scenario where you will be able to enter a few key characteristics of the project including the customer (if you have done business with the prospect before), type of project, time frame required, cost range, etc. Using these parameters, Unit4 will present you with potential reference projects, each assigned a rating of how closely they match your criteria. They do the legwork; you pick the closest, most profitable and start from there. Under the covers, Unit4 is leveraging the predictive capabilities of Microsoft Azure’s Machine Learning. This is a classic example of what is commonly referred to as “supervised machine learning” where you have known inputs (type, scope and cost range, etc. of the project) and known (expected) outcomes (cost, resources required and time frame). In this type of scenario, you have a good idea of which factors are most important in terms of predicting outcomes. These are the project characteristics you enter and you know what you are looking for – a similar project. But have you ever managed a project that looked great on paper, but in reality it was the project from hell? You can’t tell everything from the numbers. So Unit4 uses sentiment analysis to assist, which is an example of “unsupervised machine learning.”
With unsupervised learning you infer and identify patterns we might not otherwise see on our own. What is the most common word used? Perhaps you find it to be “team.” What other words or phrases are used in conjunction with team? If they are “complaints” or “excessive overtime,” perhaps the team is complaining about too much overtime. Is that sustainable? Do you really want to use this as your starting point? By automating the process, Unit4 delivers on all of the pillars of a “self-driving” ERP, from automating manual tasks to reducing input and asking only for input at the moment of action. It can sense problems, as well as potential opportunities and give intelligent recommendations.”
This is just one example from one industry. We’ll see many more of these scenarios being played out in business very soon. The full Mint Jutras report is available to download