Small and Smart: How AI is powering SMEs in India's e-commerce boom

Economic Times

19-06-2025

iStock AI has been a great help not just for sellers but for consumers as well.
Since OpenAI launched ChatGPT, a conversational chatbot powered by artificial intelligence (AI), in 2022, the world has not been the same. Although AI wasn't entirely new to businesses and consumers, ChatGPT showcased the potential for technology to be used across various sectors with diverse and nuanced applications. The world of supply chain and e-commerce had already begun employing data analytics and machine learning (ML) to gain insights into consumer behaviour and facilitate demand forecasting and inventory management. As AI continues to advance, experts say it is reshaping the e-commerce industry in India and changing the way Indians shop online.
People no longer use short keywords while using AI tools like ChatGPT for shopping online, according to Kapil Makhija, CEO & MD of Unicommerce. Instead, they ask complete, natural questions, such as 'What are the best high-quality shoes for running?' or 'Show me kurti options under Rs 1,000' or 'Show me running shoes that can be delivered in 2 days.'
'These and more are real examples of how customers are using AI in their digital commerce journeys,' says Makhija. 'Evolving user behaviour, expansion in the number of online channels, growing emphasis on enhanced user experience and the need to drive higher operational efficiencies are driving AI adoption in multiple ways,' he observes.To stay visible in AI-driven conversations, he says, product data needs to mirror the way people speak with natural phrases like 'reliable running shoes for daily jogs' or 'suitable for gym workouts', which mimic how shoppers describe their needs in AI tools. 'The more closely product information matches the way people actually describe what they're looking for, the more likely it is for the product to be recommended by conversational AI tools.''AI can enable an enhanced search experience by using natural language processing (NLP), which helps understand customers' search intent while offering conversational search. It can allow customers to refine queries through dialogues, like a discussion with a human salesperson,' he says.
Additionally, AI is helping these demands sync up with the inventory, says Makhija. 'When a customer looks for running shoes that can be delivered in two days, it will look at available data and give the results. If a brand's inventory is not synced across channels, the choice will not show up, even if it is available,' he explains.'Real-time inventory visibility is no longer an operational strength; it is the starting point in an AI-powered shopping world. While there are a growing number of sales channels like multiple marketplaces, quick-commerce chains, brand websites and apps, physical stores and omni-channel integrations that connect brands to sellers, the inventory is typically held in a few locations only. Technology helps track real-time inventory across various points of sale and storage,' he says.At the same time, he points out that AI can simplify the complexity of selling on multiple e-commerce platforms by automating product listings to ensure consistent descriptions and images across platforms.
For automating operations, e-commerce enablement platform Shiprocket has recently launched an AI-integrated model context protocol (MCP) server. This server allows AI agents to perform key operations, such as checking shipping rates or tracking orders, all through a single line of instruction.
Saahil Goel, Co-founder and CEO of Shiprocket, says that India has a very good opportunity with AI due to the abundance of unstructured data in the country. This includes different accents, languages, and voices, which means that many things that could be automated in the West remain unautomated in India.'Additionally, India has struggled with DIY (Do-It-Yourself) in the past, and it still hasn't accepted it to a great degree. I do think that can change with AI. Right now, the merchant depends on service providers for many things, all of which can come down to a prompt. Today's AI-powered co-pilots are very intelligent; they can be connected to a merchant's business, their ad account, their marketplaces, Google Analytics, etc. A merchant can simply ask the co-pilot to do a reconciliation or ask, 'How much money does Amazon owe me?' and they will get accurate answers. SMEs could have a McKinsey consultant in their pocket,' he says, adding that often SMEs struggle with not knowing, and then the cost to deliver that knowledge to them is very high in a service-orientated market.AI also helps with eliminating errors, especially in logistics, such as getting the right state and PIN codes, he says.Makhija also highlights that logistics accounts for a substantial expense in e-commerce operations. Different products, geographies, modes of payment, values of products and more determine the right courier selection for a certain order. AI models predict which courier will deliver fastest or with the least risk of RTO (Return to Origin) for a specific package or geography.'For businesses using multiple delivery partners, AI engines help allocate deliveries across different courier partners based on service levels, area efficiency, or cost. Shipway is Unicommerce's logistics management platform that provides both courier aggregation and courier automation services. Shipway's smart courier automation services help it dynamically suggest the best carrier based on current rate cards, package weight, distance, and urgency. This helps businesses reduce RTOs, improve delivery SLAs and manage costs,' he says.He adds that AI can also help in RTO (return-to-origin) reduction and fraud detection. AI helps in predictive analytics to determine likely failure chances of a delivery based on address patterns, COD risk, or past delivery success. 'AI can predict the likelihood of delivery failure and automate proactive action like calling the customer in advance. AI can also increase the efficiency of fraud detection by flagging suspicious delivery patterns, like unusually high failed deliveries from a specific area or user,' he says.AI has been a great help not just for sellers but for consumers as well. As per Praveen Govindu, Partner, Deloitte India, AI now permeates nearly every aspect of the digital retail experience. From bot-led customer support to reduced cart abandonment, AI has a lot to offer to enhance consumer experience.'Virtual assistants and chatbots now handle a wide range of customer interactions—from answering questions to guiding users through purchases and resolving post-purchase issues, improving conversion rates and post-purchase satisfaction. AI also tackles the persistent problem of cart abandonment by sending smart, personalised reminders and offering targeted discounts on abandoned cart products. Real-time chatbots also step in to resolve last-minute doubts, nudging customers to complete purchases,' he says.He continues, 'AI-driven AR (Augmented Reality) allows customers to visualise how products like cosmetics, clothing or furniture will look like on real users/set-ups, bridging the gap between online convenience and in-store confidence.'However, this technology comes with its own set of challenges, such as outdated data quality, over-automation, integration issues and privacy risks.AI can automate only what is well-defined. If business processes are flawed, automation will simply make bad outcomes happen faster. Over-automation can also lead to inefficiencies and customer frustration if not properly managed. Further, integrating AI with existing systems can be difficult, especially with legacy infrastructure causing delays and additional costs. Implementing AI involves significant upfront investment, ongoing maintenance and employee training. 'These can be barriers for small businesses,' he says. He also points out the lack of human touch.According to Deloitte India's Govindu, AI cannot think strategically, create original ideas or form emotional connections with customers. 'It only excels at well-defined and repetitive tasks.'Adds Shiprocket's Goel, saying that AI is also capable of generating fake content and fake reviews and thus has the potential to be abused. Interestingly, it is also good at catching these fake reviews and accounts and checking whether they are spam.Nevertheless, the future looks promising as per the experts. Goel believes that AI's applications will go deep into regional tier II and tier III markets using video and voice-based intelligence.'Today, when we work with brands or B2C firms, there are typically 4-5 segments—male, female, urban, non-urban, etc., and their combinations. But, in reality, there could be 1000s. You could do thousands of creatives for these thousands of people. With AI, it becomes possible to run the same ad in different versions as per every city,' he says.Talking about the future use cases, Govindu said that there will be a lot more autonomous product tagging and catalog management as well as real-time profiling and customer targeting. AI will automatically tag, categorise, and update product listing based on image and text recognition, he says.'By analysing real-time data streams like browsing patterns, cursor movements, and session interactions, AI constructs evolving customer profiles and plans hyper personalised targeting strategies. For example, AI can monitor micro interactions (scroll depth, hover duration) to infer intent adjusting recommendations mid-session. It can also analyse externally integrated data like weather, location, and device-type to tailor offers,' he says.
On the occasion of World MSME Day on June 27, 2025, ET Digital will also open registrations for its sixth edition of the ET MSME Awards 2025. The coveted award programme celebrates India's top MSMEs for their achievements and contribution to the nation's economic development.