IndiaMART InterMESH Limited, a B2B online marketplace that connects buyers and suppliers, has undergone a significant technological transformation, from directories to websites and mobile-first apps. The Noida-based firm, operating on a dual-cloud structure for several years, is ahead of the curve as organisations increasingly migrate to the cloud. In a TechCircle interview, CIO Nikhil Prabhakar highlighted the company’s recent digital innovations, including Artificial Intelligence (AI)-driven conversational commerce and optimised search, all while maintaining strong cybersecurity. IndiaMART's proactive approach to tech innovation and adaptability solidifies its leading position in the B2B marketplace. Edited excerpts:

Which areas in technology are your focus this year and why?

Our technology focus this year centres on AI-driven enhancements, scalability, and application modernization to create a seamless B2B ecosystem. We're enhancing accessibility with multilingual voice and photo search powered by AI Vision Language Models, DaViT, and AI-powered chatbots for improved customer experience. We're implementing horizontal scaling and load balancer sharding in our mobile infrastructure to efficiently manage traffic, especially during peak periods. In the areas of application modernisation, we're using Kubernetes and Docker for containerisation to ensure website scalability and resilience. We are also consolidating logic and promoting functional programming.

Any recent technology initiative that has led to an increase in Return of Investment (RoI)/customer experience or any other success?

Our initiative to fuel AI-powered conversational Commerce has not only streamlined business transactions but also reinforced trust in digital B2B commerce. The introduction of IM Insta — a seamless integration of WhatsApp with the Lead management system is a key initiative to help sellers instantly connect with buyers. This has helped in reduced follow-ups, ensured less intrusion, and increased buyer responsiveness by three folds. Further, our new photoSearch 3.0 model also reduced the average response time searching products on the platform from 4.7 seconds to just 2.5 seconds and enhanced the model's accuracy significantly over traditional models. This has ultimately improved the relevance of the search results, and reduced the margin of error, allowing us to provide the most pertinent results to our users' queries.

Considering advanced cyber-attacks in the e-commerce sector, what kind of measures has the company taken?

The increased sophistication of cyber-attacks in the e-commerce sector makes it mandatory for any company to have a robust security infrastructure in place. To address advanced cyber-attacks, we employ a multi-layered approach, including a multi-cloud strategy, strong encryption protocols, integrated web application firewalls, robust monitoring systems, continuous authentication, and strict access controls. Additionally, we follow measures that include AI-driven fraud prevention, real-time monitoring, armour threat intelligence rules integration, regular security audits, and third-party penetration testing.

How much does your team contribute to tech innovation? What’s its size/composition?

At IndiaMART, we continue to upskill, monitor the sector advancements and implement what works for us with customer eccentricity. This reflects well in our workforce, encouraging them to innovate, adopt and build solutions which makes us front runners in the B2B space. About 11-15% of our workforce actively contributes to tech innovation across AI & data, cybersecurity, cloud & scalability, vernacular/conversational commerce, and real-time analytics/automation. Our team comprises software engineers, cybersecurity experts, AI/ML specialists, data scientists, and cloud & DevOps engineers.

Are current GenAI/ agentic AI solutions providing a return on investment? What's the future? What kind of innovation are you bringing with these technologies?

Generative AI and Agentic AI are delivering RoI through improved operational efficiency, user experience, and business conversions. Fine-tuning LLMs with proprietary data is improving automated product classification accuracy to approximately 90%. Our Lead Management system leverages AI to provide tailored suggestions in Hinglish, increasing seller efficiency and engagement (over 60% of messages are sent via suggested replies). We aim to expand AI applications to optimise more seller-buyer interactions and enhance inclusivity.

On a scale of 1-10 how much would you rank the company's digital transformation efforts and why?

On digital transformation, IndiaMART rates itself highly. We were early adopters of the internet and cloud technologies, even operating on a dual-cloud structure ahead of the curve. Our expertise in interoperable technologies, modern tech stack, cloud-native solutions, microservices, and long-standing AI initiatives demonstrate our agility and innovation.

When looking at competitor benchmarking, which key technology areas do you see a need to improve?

In competitor benchmarking, cybersecurity remains a key area for improvement. We maintain a robust security infrastructure, constantly upgrading and benchmarking against sectoral advancements. Our zero-trust architecture, enhanced compliance with ISO 27001/27701, and proactive scanning of cybersecurity standards are vital.

What’s going to be your tech focus in the next 1-2 years?

In the next 1-2 years, we believe that a strong focus on the adoption of open-source technologies like Go (Golang), PostgreSQL, and microservices will be pivotal in driving innovation and scalability. It helps in building high-performance applications, while being reliable for large datasets, as yet. By embracing these open-source solutions, we can reduce costs, increase development speed, enhance system reliability, and better serve our customers in a dynamic digital landscape.