Welcome to the year of prescriptive analytics powered by AI
Posted: 10 January 2018 | By Charlie Moloney
A lot of the so-called intelligent technologies that exist today are good at deriving descriptive, diagnostic, and sometimes predictive insights.
That means that they can look at, for example, your IT system, your supply chain, or your Customer Relationship Management (CRM) and tell you what has been happening, why has it been happening, and what is likely to happen.
However, this is not a high-value add because you probably already know who you’ve been trying to sell to, whether you’re likely to get them over the finish line, and why.
Having a system that can just predict likely outcomes is not the artificial intelligence (AI) breakthrough that we have often been promised.
The future of AI is prescriptive systems, which can prescribe you effective actions.
Does prescriptive AI exist today?
We may not realise it, but many of us follow prescriptive recommendations from a machine almost every day by following the suggested route given to us by directions apps on our phones.
These apps do a fair job of making predictions about the best route based on factors like traffic and weather, so we usually follow the directions they give us without question and reach our destination around the time they predict.
Despite some high-profile mishaps involving drivers faithfully steering their cars into disaster, navigation systems are generally accepted as reliable.
For more complicated problems, and especially in a working environment, people are hesitant to follow the advice of such machines because they are not convinced that they can be trusted.
Once the accuracy and scope of prescriptive AI improves, more industries will develop tools that are as usable and reliable as directions systems to tackle their complicated day to day business problems.