Artificial intelligence (AI) has spread its wings across industries, including retail. For the longest time, this sector relied on brick-and-mortar stores, but with e-commerce making a splash, a phygital experience has become the order of the day. No wonder, retailers are working consistently to provide shoppers with experiences that are personalised, convenient and much more satisfying. But there’s a lot more that goes on behind the scenes.

Moreover, digital transformation in retail boils down to connecting things seamlessly. With reams of data available today, it is important to convert it into insights, but in an intelligent manner. This is where new-age technologies like AI and machine learning help. They don't just offer innovation and help in increasing the revenue, but also offer more value to customers. 

Jibu Elias — Reseqarch and Content Head, INDIAai had a freewheeling chat with Debjyoti Paul, President — Digital Experience, ITC Infotech, to understand the shifts in retail and what the future really looks like. 

A sneak peek into the retail landscape

Although AI has helped to make great strides in retail, Paul points out that we only laud the aspects that are visible. Instead, he turns our attention to how goods and services are being delivered today. Citing an example, Paul says that something as simple as soaps and shampoos were not easily available a few years ao, in terms of the distribution. 

“The distribution was through FMCGs, who ran their own distribution networks. Today, there are so many intermediaries, who make the process seamless, plus there’s technology, of course. The way goods reach the consumer has fundamentally changed, even in the rural parts of the country,” he explains.

In reality, the success of retail is hinged on goods distribution, on services being made available, and how payments can be made in a smooth and seamless manner with the help of AI. 

Diving deep into AI

Speaking about his organisation, Paul shares that they are a relatively small technology services organisation, as compared to many of their peers. That’s exactly why they are selective about where they play. 

“Our clients see value in us when we bring depth in our engagements, and therefore we choose the areas where we have prior experience of delivering business outcomes that are very visible. For instance, retail companies are focused on what they call revenue growth management, which comprises how they spend their trade budgets. Or how do they promote their products in the distribution system? How do they do better assortment planning? How do they do better pricing? 

Now, think of this as a country where we have, let's say, a million and a half outlets being served by many of these companies. How do you make that equivalent concept of personalization happen when each of those retail outlets are being served? Can you really, therefore deliver a very personalised assortment?,” he explains. 

Paul adds that their strength happens to be emerging or developing markets, where there's a lot of traditional trade.

“Simultaneously, in the banking space, a lot of energy right now worldwide is going into staying on top of the risk area, and therefore financial crime is the key area of focus. How do you really address the anti-money laundering requirements? That is a key area of focus,” he shares. 

What sets them apart

There’s a lot that sets Paul’s organisation apart, but what’s most important is how they are able to bring in that element of explainability, with the help of AI. 

“Our focus as a business philosophy is really to ensure that we lean towards less of a black box-oriented recommendation, and that determines what is the science that we apply and what are the models that we choose. But at the same time, fundamentally, the question of explainability is even more relevant when there are fundamental opportunities to insert bias and recommendations,” he explains. 

Elaborating with the help of an example, Paul says that when they apply AI to the recruitment process, and are screening resumes, the key is to ensure that there are no biases. But how does that happen?

“Our approach is really guided by a set of principles, which determine where we apply AI and where we need special guardrails when we are dealing with a certain set of data. So, with our customers for instance, an important checkpoint internally from an assurance point of view is that we are dealing with sensitive customer data, which can insert biases in the recommendation. That is the critical area where we invest our energies. So, if we are comfortable with that framework, we are clear that we are not doing anything that we don’t want to be seen doing,” he shares.

Implementation of Responsible AI

On the subject of responsible AI, Paul believes that the only way to ensure its implementation is to have a regulatory framework. 

“How do you use data to make a recommendation or a decision is what we were referring to? What are the moral outcomes and therefore, when do you actually implement the trigger to say – Hey machine, you’re wrong and now you need to let go of your output, right? Those things will have to eventually have a regulatory framework. And the current trajectory is that every country will go for its own set of regulations. Will we get to a common framework? There are initiatives but that will require serious work,” he says. 

Major challenges and roadblocks

Paul shares that trustworthy data availability is a big challenge, and at the same time, it is critical for the regulatory framework to be clear. 

“In India, for instance, we have a huge lack of awareness on how data will be used. When we live in a city, all our apartment complexes happily ask for so much personal data from each of us residents. Now who’s protecting it? There is no awareness or concern about it. Anybody walks in, they have to give all kinds of personal data. When we show up in offices, we are supposed to show ID proof and government-mandated ID proof and it gets recorded. How does it get used? Therefore, the use of data and the regulatory framework around that again is something that goes with it. To me, that is something that can really facilitate or curb our ability to move forward,” he explains.

Moving on to a skilling point of view, Paul adds that while we may be producing the largest pool of qualified talent from around the world, but it is only limited to certain advanced talent areas. 

“ AI is affecting every sphere of work. How do we make that available to a broad set of citizens is something we need to think about? Therefore, from a competitiveness point of view, India can really think about what the framework is and to make AI available in a generalised format, which can really take the country forward. In my mind, that really needs to be overcome,” he concludes.