Machine learning is seeping into our everyday lives, it powers our social feeds, our song recommendations, and our internet searches. It’s coming for your chatbots and the car you drive. Its potential to pervade business is obvious.
We wrote about our classification algorithm here, a tool that is forty-four thousand times faster than a human. Using machine learning to tag documents with a classification is just scratching the surface of what is possible.
The next step is for these algorithms to guide and enhance the work of humans, evaluating the work that they are doing and suggesting potential changes or improvements.
With the right training data, it becomes possible to detect anomalies and classify texts in real-time. In this scenario it becomes possible to alert users at the point that they are submitting data to a system, to ‘nudge’ them, and use behavioural economics to encourage them to adopt a different course of action.
For instance, imagine being able to alert someone publishing a tender that they need to include a requirement to reduce carbon in their specification? What about automatically coaching users who are submitting information to make sure that the data they submit is correct?
Imagine a customer services rep being asked to check if they’ve answered their customer’s question because the algorithm thinks that they’re not providing the information requested. This question might feel like an unwelcome intervention for the rep, but if it reduces the number of complaints and improves customer satisfaction, then this sort of nudging technique will become inevitable.
By intervening at critical points in any business process it becomes possible to adopt a hybrid model of machine learning, where human nuance is enhanced by algorithmic advice. Now we just have to create a machine learning model to determine whether this blog post will attract the right attention.
Talk to us about your procurement data needs, or our research and insights capability.