We used machine learning and predictive analytics to make the suit fit!

The Drop uses a combination of customer photos and self-measurements to create its suits. However, it was finding that some people had difficultly measuring themselves and those with a less than typical body shape didn't always get the right fit the first time round.

The Drop wanted to streamline this process to ensure every customer had the best shopping experience possible and had to do the minimum of measuring. With this in mind, it attended the AI Hackathon event hosted by Microsoft to try and find a solution.

But could machine learning and predictive analytics actually help to create a better custom suit?

As a trusted partner, Microsoft recommended Kalago to The Drop with the task of helping to solve this particular business problem. Over the course of the next three days we collaborated with The Drop and Microsoft to create a working solution using experimental AI.

One AI solution specifically detected anomalies in measurements. For example, if someone was unusually tall, the AI solution could detect this and flag that the customer may need to provide more manual measurements. The programme could also check that measurements taken were correct. To create this anomaly checker Kalago used Microsoft's machine learning studio and The Drop's data to create a machine learning model for anomaly detection.

Another AI prototype looked at solving customers' concerns regarding providing too many measurements and the accuracy of these measurements. The prototype was able to measure a suit just through customer photographs, removing the need for a tape measure. The team worked together on developing a model which converted photograph pixels to cms, for the tailor to work from, using a technology called Posenet.

With the help of Kalago, The Drop is currently developing these prototypes and implementing accurate working models within its web platform to streamline its business and creating a better experience for its customers.