AI monitoring and unveiling the black box

We all have heard the expression “see the forest through the trees.” It’s really about being so deep into something that you lose perspective. This happened to me recently and I thought I would share.

Let me first take a step back and say that I love my work. No, like I really love my work. I believe in the power of the technology and the building of and selling of it – I feel fortunate to be where I am. I spend hours each week deep in our technology and an equal amount of time talking about our technology trying to convince people outside how extraordinary it is. I get this stuff. But something happened to me recently – I was on a call and a prospect reacted to one of our features in a way that hadn’t happened before. In giving a demo of our unique AI Monitoring, the prospect said, “wow, that’s cool – you’ve opened up the black box.”

The black box? I had never considered our technology a black box but it occurred to me after the call that we live in a world where technology makes things happen, amazing things, but often people have no idea how it works and often it’s not easy to figure out. I had been so entrenched in my everyday, that I didn’t see the forest. 

Our AI Monitoring opens up our unique whizz-bang technology to merchandisers who want to – need to – understand how our AI machine is working to deliver the most relevant, personalized experiences. AI Monitoring was actually never meant to be used by customers, it was originally built as an in-house analytics tool where we could see how our AI machine is learning and responding at the shopper level in real-time. Today AI Monitoring is used by merchandisers who can not only transparently see the visitor’s site interactions in a session, but also how our AI is reacting to these interactions with logical probabilistic choices related to affinity scoring. 

In the spirit of transparency, here is an example that might help explain how AI Monitoring works. When you or I visit a website, we tend to browse while we shop. We use tools like ‘more like this’ or ‘you may also like’ or search to help us find what we want, or in some cases, what we didn’t know we wanted. All of this browsing behavior is an indicator of our affinity toward something. Reflektion’s AI Monitoring takes these affinity ‘scores’ and then changes the results based on these scores. So if I start viewing blue items, my blue affinity score goes up. If I switch from mens clothes to women’s, my affinity score toward women’s clothes goes up. And in both cases, the results I see as a shopper change instantaneously as I am browsing. This is the ‘black box’ that we now have made fully transparent to our merchandisers.

AI Monitor

What occurred to me when I heard the black box comment was that this deep level of AI transparency offered is very unique in the market. The lesson for me is to continue to look for opportunities to show the value of the forest as it relates to Reflektion’s technology and try to look less at the trees. 

Will Tafel