Tech Deep Dive: Artificial Intelligence (AI)

Before I dive in here it is important to baseline on what Artificial Intelligence is… The standard definition is that AI is the intelligence exhibited by machines or software.  Key ingredients to enabling these entities to “think” is that they need to be trained through machine learning on what is “good” data to use in making the predictions and what data should be disregarded.  The ironic part of this is many times humans are used to check the learning and make sure the AI program is being trained correctly. 

In Retail there are some companies starting to deploy AI based solutions to drive business value.  One solution I really think stands out is from CB4.  This company has been working for seven years on perfecting their algorithms that look for anomalies in similar situated stores for items at the SKU level.  They can then recommend to the store manager- go check “King size Snickers” as that store is under performing sales and report back (check) the algorithm as to what the issue was (example- misplaced stock or needed restocking from back room).  What I like about their solution is it integrates with the business process to drive action at the store level and hence immediate business value.  In addition, the tool has been getting 95% compliance to completing the insights from their user base.  Already the platform has delivered over one million actionable insights and the companies using it typically see .8-3% same store revenue uplift.   I also like that it is different than other solutions in that it looks at the individual SKU level- without getting to the customer/market basket analysis which may run into data privacy concerns.  Retailers that have deployed this solution have already seen up to 500% payback on the solution.  Because the insights are different at each store it is necessary to deploy network-wide across a retailer instead of only applying to a few flagship stores.  CB4 will be at the NACS (National Assoc of Convenience Stores) show in Las Vegas during October so you can see a demo or meet with their people there!

              In the future there will be more AI based solutions deploying in retail.  Key to a great solution is clean/useful data, a lot of data (to get the machine learning honed in) and lastly clear outcomes on what to do based on the data.  Over time as technologies become more main stream they are no longer considered Artificial Intelligence.  A great example of this is Optical Character Recognition (OCR)- which used to be rare/considered AI but is now common place and not considered part of the AI skill set.