AI in customer analytics: Sharpening individual focus

Illustration: Sameer Pawar

The choices we make, shopping online or using digital payments at physical stores, can tell the retailers, merchants and banks a lot about us. And today, AI programmes are sophisticated enough to be able to crunch that data and tell us much more effectively than traditional regression analyses, says Priyanshu Mishra, product leader at Crayon Data, a big data and AI company.

Crayon is a Singapore-based company that Suresh Shankar, a repeat entrepreneur, started about nine years ago, with much of its product development done out of Chennai. Crayon’s flagship product Maya.ai helps large banks, restaurants and hotel chains engage better with their end-consumers.

“The way they are looking to use AI is around the ability to understand and scale with their customers,” says Vidhyashankar Sriram, vice president for client solutions at Crayon. “How do I build scale around the intelligence that I carry?” Another area is about making their own operations smarter—for example, by being able to predict a potential breakdown in the IT systems.

Traditional methods essentially superimpose the behaviour of one segment of customers on another segment, Mishra says. At Crayon, with Maya.ai, “all of our algorithms focus on building out a ‘taste print’ or ‘taste profile’, which is essentially a combination of all the different variables”, he says. And Maya.ai tracks as many as 30 variables to define just one single individual.

This cuts across not just a person’s preferences in fashion, cuisine, and even colours, but also her behaviour itself—does she save at the beginning of the month to splurge towards the end of the month or the other way round; does she prefer a particular card for one category of purchases and maybe a wallet for a different set of purchases; does she travel frequently between various locations, which means her spends are distributed and so on.