Automating with intelligence
By Harry Worsnop, chief commercial officer at Razor
Last week I talked about how automating repetitive processes could improve the efficiency and effectiveness of the people in your business. But what happens when the processes we want to automate require intelligence: the ability to learn and apply that learning?
Machine learning
The words ‘machine learning’ tend to elicit a range of responses from mild confusion to a roll of the eyes accompanied with “well that’d be lovely, but we’re not Google”. So, allow me to be the bearer of good news: you really don’t have to be.
Most people’s understanding of how machine learning can be applied in the real world tends to centre around how some of the major ecommerce retailers are able to provide increasingly more accurate suggestions to their customers, or how Rolls-Royce use data from their jet engines to predict maintenance requirements.
These same concepts can be used in any organisation and over the last couple of years, it’s become even more feasible to do so. Microsoft Azure ML and IBM Watson are just a couple of services offering machine learning capabilities that are both relevant and accessible to a wide range of businesses, not just the few at the top.
Opportunities in every sector
Plenty of businesses have significant amounts of data that they’re doing nothing with. Some are simply neglecting the opportunity to capture any data at all. All of this data that could be hiding insights that would lead to improvements in customer service and efficiency or reductions in cost. Consider a few examples:
● Using data on frequency of use, ambient temperature ranges and the age of a piece of equipment to predict when it will need to be serviced or when it is likely to fail.
● Using data on buying patterns, stock levels, price changes and lead times to automate the purchasing of new stock – what to order, how many, where from, and exactly when to place the order.
● Using data on weather, traffic and local events to predict busy periods in a bar or restaurant.
And remember that all of this can change in real time as the systems continue to learn and refine.
Start small
We’ve written about machine learning and the concept of ‘intelligence amplified’ (using computers to help humans, rather than replace them) over on our blog, and one of the most important pieces of advice we give people is to start small. It’s not only the machine that will be learning, you will be too. You’ll need to establish what data you need to capture, how you’ll access it, where it’ll be stored and crucially, how to frame your question so that the most appropriate predictive models can be applied.
