About

I am an Associate Professor of Marketing at the Indian Institute of Management Calcutta, and I work as an AI and machine learning consultant. I study how data, regulation and digital platforms shape what consumers and firms do. I also build the systems I teach about, and I help organisations decide where AI belongs in their work and where it does not.

What I do now

My time divides across three kinds of work.

I teach marketing and business analytics, AI and machine learning applications, marketing research and digital transformation, in the MBA, executive MBA, PGDBA and doctoral programmes at IIM Calcutta. I also direct executive programmes, among them the Advanced Programme on AI for Leaders.

I do research, mostly on data-driven marketing, business analytics and consumer welfare in emerging markets.

And I consult. I advise government bodies, large corporations and technology start-ups on AI and analytics adoption, digital marketing, pricing and revenue management, and data-driven operations. A good part of that work is practical: small language models for routine tasks, agent systems that run a defined process with a human at the decisive step, and the less glamorous work of redesigning a process so that the tool changes something.

The career arc

I did not start in a business school. I trained as a mechanical engineer, with a B.E. from PSG College of Technology, and my first job was in a technical role at Tata Consultancy Services in India.

I then went to Cardiff Business School in the United Kingdom for an MBA, which I completed as the University Gold Medalist. After it I worked in the UK in financial services and publishing, in roles that covered marketing analytics, digital marketing and marketing technology. That was where I learned what it is like to be the person responsible for a number that other people rely on.

I came back to India for a PhD in Marketing at the Indian Institute of Management Ahmedabad. Teaching and research followed. I joined the Indian Institute of Management Udaipur as an Assistant Professor in 2015, moved to IIM Calcutta in 2016, and have been an Associate Professor there since March 2023.

The arc is less of a detour than it looks. Engineering gave me a habit of asking how a thing is built. The industry years gave me a respect for operations. The PhD gave me a method for deciding whether something is true.

Research

I have 27 published papers, including work in the Journal of Marketing, which is on both the FT50 and UTD24 lists. My Google Scholar profile shows 836 citations and an h-index of 14.

The themes recur: pharmaceutical price regulation and access to medicines; conspicuous consumption and income inequality; marketplace literacy at the base of the pyramid; ICT-enabled development; and, more recently, how consumers respond to AI and digital assets. The methods are field experiments, quasi-experiments, event studies and choice modelling. My work has been funded by the Stigler Center at Chicago Booth, the Department of Science and Technology, the Global Development Network and the Ministry of Housing and Urban Affairs, and I have received the PRME Faculty Excellence in Research Award.

How the research and the building feed each other

The two activities correct each other.

Research taught me to be suspicious of a result that arrives without a comparison. Much of my work asks what happened because of a policy or a platform, as distinct from what happened alongside it. When I write about AI, the same habit shows up. If a vendor reports a gain, I ask what it was measured against. If an ad platform reports its own conversions, I ask for a holdout. It is the same question in a different setting.

Building taught me what papers rarely say. A model that works on a benchmark can fail on a firm’s own data. An agent that is technically sound can stall because nobody owns the process it changes. I have built a marketing system whose agents spend within a daily limit and ask above it, and a retail operations system with about 50 agents for RR Mart, a rural supermarket chain where I sit on the board. I have also fine-tuned a small model for a scientific publishing services firm: on 94 held-out queries, it classified 88.3% correctly, against 63.8% for a prompt-only baseline. I trust numbers like these more than assertions, which is why I report them with the test they came from.

What I learn from these engagements goes back into my teaching cases and executive education, so that the classroom stays close to live practice.

How I work with organisations

I start with the process, not the tool. Before anyone chooses a model, I like to see how the work is actually done: who does what, who approves, where it stalls. I would rather spend the first weeks with the people who do the work than with a vendor’s demonstration.

I use the cheapest model that does the job well enough, and I test that claim on the organisation’s own examples. For high-volume, repetitive tasks that often means a small model. For decisions where accuracy matters most, it means a stronger model and a person.

I put limits in writing. Where an agent spends money or speaks for the firm, I set the threshold above which it must ask, and I name who may say yes.

I report what I measured and what I did not. Where I do not have evidence, I say it is my opinion.

Past and current engagements include the Ministry of Housing and Urban Affairs, RR Mart, ThinkRisk, Edelweiss, Dr. Reddy’s Laboratories and NCHMCT, and I am a strategic advisor to MetaDecrypt AI Atlas.

Closing

I believe knowledge matters most when it empowers others, whether they are students, firms or communities. If you are working out where AI fits in your organisation, or what to test before you spend, I would be glad to hear from you.