Why AI is bad at innovation

Markets change over time. The past is not the same as the future. This might seem obvious but organisations spend a lot of time analysing data that came from the past (e.g. sales) to then act on what to do in the future.

AI is the great average machine - literally predicting what to say based on the most probable answer. Take user feedback for example, it can give you back the common themes for a particular group. This is not great for innovation because we are looking for signs of the shifting marketplace. We are looking for the canary in the coal mine.

These insights come from the outliers - the things that are not common. The small subtleties that suggest something different is happening. For this we want the “not-average” machine, the machine that can look for the start of what will become a trend.

Having AI give averages of feedback is not a bad thing - it can help you refine your product. But this is an efficiency play, you cannot grow in business forever just making making an old product continually refined. Needs and wants change and you need to adapt if you want to grow. The more certain and algorithmic a marketplace, the easier it is for a competitor to enter, potentially commoditising your product offering. You need a system for determining “what’s next”.