For Dinesh Agarwal, Founder & CEO, IndiaMart Intermesh, data has never been an afterthought, it has always been the foundation. Long before AI became the defining theme of enterprise conversations, IndiaMart was quietly building one of the largest data ecosystems in the country’s digital economy.
“IndiaMart has always been about data,” Agarwal says. Over nearly 30 years, the platform has accumulated data from close to 80 lakh suppliers and over 20 crore registered buyers, spanning nearly 1,00,000 micro-categories of products and services. These interactions originate from thousands of towns and cities across India, creating a uniquely deep and diverse dataset that few platforms can match.
Beyond listings and searches, IndiaMart sits on decades of buyer–seller interaction data, calls, messages, enquiries, and negotiations, flowing through its end-to-end lead management platform. “All the search data, all the buyer–seller interaction data, has been stored with us for a very long period of time,” Agarwal notes.
In the pre-AI era, much of this information could only be analysed at a surface level. That has changed decisively with the rise of LLMs and multimodal AI. IndiaMart today is unlocking value from millions of PDFs, both text-based and scanned, that were earlier lying unused. “Now, with multimodal models, we can read PDFs, summarise them, extract tables, and generate insights. That is helping us dramatically improve our category intelligence,” he explains.
The real opportunity, however, lies in understanding how commerce actually happens across India. By analysing buyer–seller interactions at scale, IndiaMart can now infer pricing trends across products, geographies, and demand cycles. “We can understand what kind of prices are being traded for what goods in which city,” Agarwal says, adding that this could eventually lead to the creation of a modern commodity exchange layer on top of the platform.
Customer engagement is another area where AI is beginning to reshape the marketplace. IndiaMart is increasingly relying on AI-powered voice agents to support buyers and sellers round the clock. These agents can converse naturally, switch languages mid-conversation, and operate across an almost infinite range of product categories. “No human can be perfect across 1,00,000 categories,” Agarwal admits. “One moment you are talking about medicines, the next about chemicals or electronics. Training people for that scale was always difficult.”
KKMs have changed the equation. “They have almost read everything. You can talk to them about medicine, neurology, physics, or electronics, in any language, at any time,” he says. While Hindi and English are currently the primary customer-facing languages, the underlying capability extends to multiple Indian languages. Guardrails are being carefully audited before wider deployment, but Agarwal believes the shift is inevitable. “By the end of this year, a very large part of customer interaction will be voice-AI driven.”
IndiaMart’s ability to adopt AI is rooted in the platform’s evolutionary architecture. When the company started operations in the mid-1990s, digital commerce was barely a concept. “I didn’t know how big this would become or what kind of architecture would be needed,” Agarwal reflects. The platform began as an export–import marketplace driven largely by email. As mobile phones became the primary identity layer in the mid-2000s, IndiaMART pivoted accordingly. Later came the shift from desktop to mobile-first usage. “It just evolved with market needs. Agility has always been the key,” he says.
Internally, AI is now being applied far beyond customer experience. IndiaMart’s category management teams, responsible for defining specifications, buyer requirements, and matchmaking logic, are using AI to compress timelines dramatically. “What earlier took years, we are now able to do in weeks,” Agarwal says. AI has found use across buyer journeys, seller enablement, and even HR functions, giving it what he describes as a “360-degree application” across the organisation.
While AI has not yet led to any significant reduction in headcount, Agarwal is candid about how roles will evolve. “People who do repetitive work, whether on the phone or behind a computer, will be reimagined,” he says, drawing parallels with the disappearance of clerical jobs after the advent of computers. However, he remains optimistic about human adaptability. “We always find new ways to keep ourselves relevant.”
On the broader challenge of AI adoption, Agarwal believes many organisations are still in the experimentation phase. “People overestimate the impact in the short run and underestimate it in the long run,” he says. For IndiaMart, AI investments are evaluated on clear business outcomes. “Does it improve quality? Does it do it faster? If not, there is no point deploying AI for AI’s sake.”
Among these criteria, quality takes precedence. “For us, better comes first, faster comes second, cheaper comes last,” Agarwal explains. “Being a profitable company, cost is not the primary driver.”
Looking ahead, IndiaMart’s roadmap includes extending AI capabilities directly to small and medium businesses. The company is exploring AI-powered CRM tools, voice agents, and chatbots that SMEs themselves can use. “We want to extend these benefits to millions of SMEs,” Agarwal says.
Perhaps the most visible change will be in how users interact with the platform itself. IndiaMart is preparing to introduce a more conversational interface, one that helps users plan, not just search. “Today, you need to know what you want,” Agarwal observes. “Tomorrow, the platform should help you plan, whether you are setting up a factory or figuring out procurement.”
For a marketplace that began its journey before digital commerce had a name, IndiaMart’s AI-driven reinvention underscores a larger truth, which is, platforms built patiently on data are often the ones best positioned to lead the next technological shift.