What consumers say vs what they buy: How AI is revealing the gap in India's shopping behaviour

Traditional consumer surveys have long helped companies assess demand, test products and understand customer preferences. However, what consumers claim they would buy does not always translate into actual purchases, particularly when they encounter competing brands, different price points and promotional offers.

This gap between consumer intention and behaviour is emerging as an important focus for AI-led market research platforms.
Why consumer surveys may not tell the whole story
PulseAI Research, an AI-powered consumer intelligence platform developed by Smytten, is combining conventional survey findings with behavioural signals to understand how consumers discover, evaluate, try and purchase products.

The platform has crossed 200 enterprise clients in just over two years and now contributes nearly a quarter of Smytten's group-level revenue.

According to the company, consumers frequently respond positively to product concepts in isolation but make different decisions when presented with familiar brands, alternative prices or competing product claims.

“The future is not survey data versus behavioural data. It is about understanding what people say, what they do, and why the two sometimes differ,” said Swagat Sarangi, Co-founder, PulseAI Research and Smytten.

How AI tracks the consumer journey

Rather than relying exclusively on questionnaires, PulseAI examines consumer interactions across product discovery, trial, shopping and repeat engagement.

For instance, researchers can introduce a mock product into a discovery environment alongside competing alternatives. They can then analyse product-page visits, add-to-cart activity and consumer feedback to identify factors influencing purchase decisions.

However, the company distinguishes between browsing, expressing purchase intent, adding products to a cart and completing transactions, recognising that each represents a different level of consumer commitment.

The platform uses statistical models, machine-learning algorithms and AI-assisted interpretation to identify patterns in consumer behaviour.

“We don’t rely on the LLM to invent the mathematics; we use it to make the mathematics conversational,” Sarangi added.

What this means for brands

For FMCG, beauty, personal-care and other consumer-facing businesses, such insights could help identify why promising products fail to convert consumer interest into sales.

Companies could potentially use these findings to refine pricing, packaging, advertising claims and product positioning before commercial launches.

PulseAI also incorporates researcher-led validation and respondent quality checks to assess the reliability of findings.

As brands increasingly seek faster consumer intelligence, combining what shoppers say with what they actually do could become an important tool for reducing uncertainty in marketing and product development.