As technology rapidly advances, artificial intelligence – and more specifically, machine learning – is becoming increasingly relevant for SparkOptimus and its clients.
Before starting at SparkOptimus in the summer, I developed various machine learning tools, ranging from a web scraper to an automated trading algorithm, a predictor for real estate prices and all sorts in between. SparkOptimus has already been helping clients for a while to define the ways in which machine learning could add value to business and I’m thrilled help develop and implement those solutions in 2019.
At SparkOptimus, my aspiration for next year is to help develop machine learning (ML) solutions in two specific domains: eCommerce and manufacturing. For eCommerce players, it’s all about selling more, and knowing why you are selling more. Useful tools for eCommerce might be a recommendation engine that suggests suitable products, or a churn predictor to increase CLV, for example. In manufacturing, on the other hand, clients face very different problems, such as how to optimize throughput and lower costs. Here, my aim is to design the likes of a predictive maintenance engine that can prevent unforeseen breakdowns and reduced throughput and instead facilitate timely, less costly repairs, as well as a demand prediction system that enables the careful matching of supply to demand.
It goes without saying that the quality of the output is 100% dependent on the input: data. Companies are collecting more and more data, believing it will magically turn into golden eggs. The most common pitfall of all, is maintaining the belief that data alone will tell us everything. To my mind, the challenge businesses will face lie largely in 1) determining which data to store, 2) labelling and cleaning data, and 3) how to connect and interlink private and public data sources.
Machine learning is, however, only useful if it is easy to use. When you consider that Alexa and Google Home are already capable of answering questions like “What was my revenue growth in the last 5 years?”, you can wonder how long it will be before CEOs can ask “How should I adapt my supply chain to save costs?”. Can you imagine the impact this will have on the consulting industry?
With machine learnings more and more readily and easily available, it is also becoming less of an “optional extra” for businesses. Just imagine machine learning opportunities are the sea, and your business is the shore. There are two choices: become a ship and ride the waves of immense value and margin, or stay put and let the water steadily erode your margin until there’s nothing left.
2019 will mark a step change in machine learning: there will be no neutral position. You either sink or you swim.
