Shipping is a massive global industry, that is not much of news, about 70% of global trade moves in vessels. One would expect then that the technology that supports the industry would be in line with the industrial best practices, but the reality is, in fact, is the opposite. Trends in technology in shipping are many times focused on gimmicks and older concepts. The flavor of the month has been blockchain and Artificial Intelligence (AI) for a while now and a lot of attention has been routed towards these technologies.
And yet, at the same time that our professional services team is asked about what products we have in blockchain and AI, which are without a doubt cutting edge technologies, the reality remains that 99% of shipping companies use tools that are 20 and 30 years old and are pushed by traditional software vendors to think that that dichotomy is reasonable. Mountains of paper on desks, invoices that are stamped on paper for payment approval, excel used to manage projects, excel and word forms that are sent from ship to shore, these are all chronic disease symptoms that indicate a very weak management cultures. A basic ability, if any, to mine data in a serious way to identify proper cause of failures, and a complete inability to really predict failure, from all those masses of data lying in cabinets and on scanned pieces of papers.
While the potential of blockchain in shipping is not quite clear yet, but seems miniscule. Bills of Lading and international transactions that are a large part of the commercial side of shipping do naturally fit into many blockchain applications. At the same time, issuing certificates by a class society in blockchain is a gimmick that has solutions that are far cheaper and easier to implement. It is quite clear that there are nearly zero real use cases to Blockchain in shipping, yet the conferences and panels are packed with talking heads and experts on the topic.
This is not the case in the world of Artificial Intelligence ("AI") or in its more relevant form Machine Learning ("ML"). The simplest way to understand ML is basically to teach a computer system to create an intuition of a 45 year old Superintendent that looks at a simple mail or a breakdown, and from his stomach identifies a much bigger issue that was not obvious to others. The complexity of shipping lends itself well to ML, but there are some significant pitfalls to watch from.
In order to even begin to contemplate Business Intelligence (BI) and machine learning in your company, to "walk before you do inverted flat spins", you need the basic building block, which is a single database ERP system. It has to be in the cloud, because the power of machine learning (AI) is in the amount of data the machine can analyze and detect trends, and the data of one company will not yield any meaningful statistics, the power of machine learning can only be realized if data of many companies can be used together, comparatively (without Personally Identifiable Information "PII" ). Once you have a cloud - single database ERP, and have basically eliminated in full all paper, scanned forms, and all data gathering systems that don't log and index that data into a centralized database, only then, you can start to consider ML in a meaningful way.
Comments