Leveraging AI & analytics to build the shoppers of future

CIOL Bureau
New Update
AI trends

Imagine a traditional retail outlet not too long ago. Even in the age of modern retail organizations, shopping used to be an uncomplicated thing just a few decades back. There were standard channels of purchasing goods and services.


Retail was clearly structured with well-defined roles and responsibilities for various functions that took care of the main areas like marketing, procurement, merchandising, supply chain, and corporate functions. Cut to the present, and we see a drastic change in the way people make purchasing decisions, payments, identify with brands, make returns, order and take delivery, and influence others in the whole cycle.

Hybrid models have sprung up all over the retail map with a plethora of offerings and models spanning the spectrum of the buying process.

Take, for instance, how shoppers are equally comfortable buying online and picking up in stores or even combining curb-side pick-up stops with a quick hop into the physical stores while they are there. Shoppers have apps and assistants that compare prices, giving them the best deals in near real-time and access to products and brands they want to consider.


Brand loyalty itself is being tested to the limit when there is information available everywhere, and there are plenty of choices for the consumer.

Issac Mathew, Senior Director, Technology for the Data and Computational Intelligence, Lowe's India Issac Mathew, Senior Director, Technology for the Data and Computational Intelligence, Lowe's India

So what will be the future? AI systems, using the collective wisdom of the people on the ground. Experience adding on to deep learning, reinforcement learning, simulators that could be used by the business or Jarvis for the store manager.


The shopper is in for a pleasant surprise – hyper-personalized offers that change dynamically based on the response, influencers, better payment, delivery and returns, real-time customization, predictive purchase assistance.

What we have done till now is an example of using historical inputs, current trends, and extrapolating to the future state based on what we know is feasible. This is conceptually similar to what AI/ML systems have done in the retail space and continue to push the envelope in terms of current and future shoppers.

Data is available, information is relatively scarce, wisdom needs to be extracted – no pun intended, but it is equally painful in some organizations that are not ready. The shoppers of the future will be happy to switch to ones that leverage data and cater to their needs in ways unheard of in the previous model of retail. Really, how long can you resist the pitch of an omniscient being that tailor offers to you as a person instead of a segment or cluster?


If the pundits are to be believed, AI in the future will be able to understand you even better than yourself.

What is already on the ground, in production, and already creating moats around customer bases are broadly around systems that look at

– Anything that improves sales


– Or optimizes operations to squeeze out the last penny and gain a competitive advantage in every transaction

– Foundational systems for data or model building (Force multipliers, not bare necessities)

Suppose we work backward from the consumer of the future: in that case, most of the forces that drive innovation will be derived from the consumers’ need to have a faster shopping experience, gives them better prices, improved selection, enables a hassle-free experience, gives them good brands that they identify with and will allow them to send social signals.


AI and ML have already been making strides in these aspects, and it is logical to expect them to grow in ways that will serve these needs:

• Faster – forecasting, network flow optimization, warehouse picking optimization, floating warehouse locations, flexible pick locations, joint AI ML models with vendor systems, flexible lead times,

• Better prices – Demand sensing and shaping, optimized sourcing, logistics optimization with dynamic fleet and channel pricing, routing options,


• Improved selection – forecasting of trends, demand quantities, mix; having a simulator as a co-worker for merchandising operations to help decide what to carry, how much and what to discard, lower out-of-stock

• Enables a hassle-free experience – Improved recommendations and substitute offers, smooth buying experience, hybrid delivery options especially for newer digital offerings, preventive reduction of returns, hyper-optimized returns processes, prevention of fraud and personal data leakage, digital activation or unlock of services/goods and delivery; security will span the phygital world with additional emphasis on prevention. goods and associated service bundles will continue to evolve, as will modes of non-standard payments, leading to additional worries on the security side.

• Good brands / send social signals – the choice and presence of brands will be more pervasive and across channels than ever, spreading messages will take on highly precise and custom message formats that appeal to individual tastes, digital cults can take on the role collectively from individual influencers, the mode, frequency and message could be influenced by AI systems that know the shopper better than themselves, social and status signals could be amplified via the combined effect of multiple channels that then act in turn to influence other shoppers. Messages could have fluid content based on geolocation, time, and seasonality. What message appeals and especially avoiding unacceptable, even inadvertently, will be increasingly decided by AI/ML models.

While it is tempting to paint the flying cars scenario, it is also important to identify the trends that may not make it. They are usually the ones that sound cool but aren’t helpful. If we look at a metric that measures the usefulness to coolness ratio, the low ranking ones take too much effort for developing such systems compared to the impact, and they fall along the way, except in some GitHub repositories where they may be used to seed future ideas that find the right usefulness to coolness ratio or will be discarded entirely as fossils.

Like the Renaissance disrupted the world with sponsors for art, business sponsors most of these AI/ML efforts. The ones that ultimately benefit the ecosystem of shoppers and the company will be naturally pruned to give rise to better systems, with more sponsorship and chances of development into further stages, a feedback loop with amplifying effect that is self-sustaining. Together with the market forces, this will determine the future of shopping, and this is definitely the space to watch.

Author: Issac Mathew, Senior Director, Technology for the Data and Computational Intelligence, Lowe's India