Digital Transformation Isn’t About ERP Anymore, It’s About Automation With Machine Learning

As long as ERP-based automation is confined to the automation of processes embedded in ERP applications, the ERP world will lag behind next generation digital transformation.

Pre-Covid-19 Digital Transformation 

Not so many years ago companies defined “digital transformation” as “the use of new, fast and frequently changing digital technology to solve problems.  It is about transforming processes that (are) non digital or manual to digital processes.” Companies rushed to standardize their business processes with enterprise resource planning (ERP) systems, like the formidable, almost pervasive, applications from SAP and other ERP vendors.  The value proposition was – and remains – enterprise centralization, standardization and integration across internal and external business processes, cost savings and – in spite of their monolithic architectures – some flexibility (compared to the chaos that existed before).  Kim O’Shaughnessy says it better:

“Enterprise resource planning (ERP) systems are used by organizations looking to manage their business functions within a centralized and integrated system.  ERP is commonly used by companies working within the supply chain to help keep track of all the moving parts of manufacturing and distribution.  However, ERP can be utilized by a number of different industries including those in healthcare, nonprofit groups, construction and hospitality.  Organization needing to manage their staff, customers and inventory can all rely on ERP benefits … ERP stores all entered data into a single database, allowing all departments to work with the same information. Additionally, all this data can be organized, analyzed and made into reports.  ERP brings together customer management, human resources, business intelligence, financial management, inventory and supply chain capabilities into one system.”

For many companies these capabilities were – and remain – a God-send.  The basic functionality of enterprise applications can organize a company’s business processes and whole business models.  All good, for sure.  But is this a sustainable role given what’s happening with machine learning?  Clearly, companies will invest heavily in automaton to save money and eliminate, streamline and optimize processes.  They will map and mine their business processes for maximum impact.  They will do this as quickly as possible.  The Covid-19 pandemic has obviously accelerated the drive toward automation. But the sprint to automation was well underway prior to the pandemic.

Digital Transformation is Now Digital Automation

Digital transformation today is still about organization and standardization, but it’s also about automation.  In fact, post-Covid-19, it will be much more about automation than functional standardization.  While enterprise applications vendors (like SAP) and ERP vendor enablers (like UiPath) are investing heavily in automation, the most automated companies will move past their enterprise applications to functionality that’s increasingly automated outside of older application architectures. Why is this important?  For one thing, post-Covid-19 companies will focus much more on the processes that can be automated than ones than can be “tracked” and “managed.” Ideally, much of what enterprise applications “users” do will be replaced by RPA and other intelligent bots.  The obsession with “UI” and “UX” (user interfaces and user experiences), for example, will eventually vanish altogether.  

How many companies will deploy new (or migrate from existing) ERP systems over the next ten years?  Not as many as did ten years ago, that’s for sure.  Instead, they will:

1. Identify and model business problems
2. Map and optimize business processes (via process mining) that identify process optimization opportunities
3. Identify, collect, validate and leverage structured and unstructured data onto target business processes
4. Match problems, processes, data and machine learning algorithms
5. Extend the approach to as many processes as possible

These activities are all focused on automation.  Said differently, they’re focused on automating — and replacing — ERP modules and the processes like “customer management, human resources, business intelligence, financial management, inventory and supply chain capabilities.”  In fact, much of the internal and external business cycle can be automated.  But automation extends beyond tactical goals.  What are the best profitability strategies?  What new markets make the most sense?  How should we innovate?  When we increase the range of algorithmic applications across supervised to unsupervised machine learning, we can move from tactical to strategic automation.

Just like almost everything, all of this will end up in the cloud.  In time, intelligent companies will automate fully in the cloud.  Note that the major cloud providers – AWS, Azure, Google and IBM – already offer robotic process automation (RPA) services of one kind or another.  There will be more.

Post-Covid-19 – and forever – companies will shift their platform thinking from large centralized, standardized and integrated enterprise applications to platforms that automate business processes and models independent of a single application.  While ERP vendors rush to automate their functions, other vendors will take a very different approach.  They will identify and model business problems; map and optimize business processes (via process mining) well beyond the processes embedded in ERP applications; identify, collect, validate and leverage structured and unstructured data; match problems, processes, data and machine learning algorithms; and extend the approach to as many known, new and anticipated processes as possible.  They will pursue all this outside the boundaries and constraints of their ERP applications. As long as ERP-based “automation” is confined to the automation of processes embedded in the ERP application itself, the ERP world will lag behind next generation digital transformation as it transforms to digital automation.