Intelligence As A Service
August 18, 2016
This time a year ago, Artificial Intelligence (AI) was something fun to think about but seemed to be far from a realty for most of retail. Companies like IBM with their Watson was cool to see in action and some of the concepts they came up with were very cool too. But, again, seemed to be out of reach for the rest of us.
Why? Because most of us, who have been around the block a few times in retail technology are frequently reminded of all the data we capture and how little we actually do with it to drive intelligence related decisions. Yes, we are pretty good at developing algorithms for predicting certain purchases as well as for planning and distribution but to think along with a customer, give them what they want, then continuously learn, not so much.
Most of us old-school retail technologists are simply too ‘puristic’ in our thinking too. We have been taught how important it is to own all the data along with all the related contexts for the data. In other words, we would have to build our own ‘data lakes’ (subject for another day) and then begin winding our way around, in and through it. I don’t know about you but when I am faced with a project with a bi-zillion moving parts, I get overwhelmed to the point of paralysis and that is just what all that data feels like.
Furthermore, this would require additional resources such as data scientists, new analysis tools, the ability to extract data outside of our own infrastructure and place it in context. This could take years before we would see any real effects, not to mention the cost.
But then I read an article about 1-800-Flowers and their partnership with IBM’s Watson. At first I was very critical because I saw the use of AI to help someone buy flowers a waste of intelligence. Predictive analytics would likely do you just as well and for a lot less effort, not to mention cost. Next I read about Macy’s partnering with IBM’s Watson and it got my attention. If these folks could use Watsons engine and its head start into intelligence, AI may actually BE right around the corner. IBM has placed Watson on the cloud and though the use of an API layer placed between your applications and Watson, AI is possible. Many retailers use API layers so the concept is not new. Also, it is efficient and doesn’t require a team of data scientists.
Intelligence-as-a-service! Sounds like just the answer.