Technologies used by the Port of Hamburg make it possible to digitize many business processes, which is key to creating a Live Business – one that can coordinate multiple functions to respond to and even anticipate customer demand at any moment.
1. Edge algorithms, such as speech or image recognition, make decisions based primarily on their ability to interpret input with precision and then deliver a result in real time.
2. Server-level algorithms gather data from edge algorithms and report on both the results of data analysis and the analytical process itself. For example, the complex systems that generate credit scores also tell applicants how they can raise their score in the future.
Successfully digitizing a business process requires you to build a model of the business process based on existing data. For example, a bank can use predictive analytics to try to determine what existing customers might do next.
When the results of a predictive model start to drift significantly from expectations, it’s time to examine the model. You need to determine whether you should dump old data or include new variables, such as marital status, that further refine your results.
Data shouldn’t always be perfect. For example, to train an optical character recognition system to recognize and read handwriting in real time, your samples need to include sloppy scrawls so the system can learn to decode them.
When machines are talking to each other, humans only slow things down. In human-facing interactions, though, it’s still best to digitize the part of the process that generates decisions, but to leave it to a human to check and finalize the decision.