Companies in manufacturing or any other other sector, that were early adopters of machine learning and AI, usually started their digital transformation journey by hiring a team of data scientists. To their disappointment, projects took a lot longer than expected. To make things worse, the return on investment was rarely positive. The data scientist was not to blame however: legacy systems and bad data prevented the data scientists from doing their job.
Since then, companies have learned from their mistakes and understand that digital transformation begins first and foremost with modernising their data architecture. This gave rise to the data engineering profession tasked with building scalable, reliable and maintainable data infrastructure to empower the data scientists to do his work efficiently.
However, this makes digital transformation a chicken-and-egg story: you need to invest a lot of money before you can start generating or saving money with machine learning. It’s really hard to convince upper management to grant the IT-department 5 million dollars and three years time to build a proper data infrastructure, before any value is delivered. Especially if your company is not an IT-company. What if this undertaking fails with nothing to show for it?
Here’s the thing:
A good digital transformation strategy will likely follow a middle-of-the-road approach. The goal is to keep building the ultimate data architecture while incrementally delivering value through data use cases. These use cases could be as simple as calculating an important metric to discuss in daily production meetings. . The periodically generated cashflow or savings keep the company’s management happy to continue financing the long term objective.