AI Chat Becomes the New Starting Point for Online Shopping, Southeast Asian Retailers Face Data Challenges

The fourth edition of Salesforce's 'State of Commerce Report,' released in October 2026, shows a 200% year-over-year increase in the use of agentic search as the first step in the shopping journey. The survey covered 3,450 business professionals across 13 industries in 20 countries, including 100 respondents from Singapore, and also gathered feedback from 4,689 consumers, along with behavioral data from over 1.5 billion shoppers in 37 countries. During each statistical quarter, e-commerce traffic driven by AI chat grew between 150% and 428% year-over-year, while overall e-commerce traffic saw only single-digit to low-double-digit growth during the same period. From August 2025 to May 2026, the share of consumers discovering products through brand-owned channels fell by 7%, traditional search channels dropped by 15%, while new discovery channels, including AI assistants, social media AI, and delivery apps, grew by 38%. The Southeast Asian e-commerce environment is already highly fragmented, with consumers in cities like Singapore, Jakarta, Manila, and Bangkok often switching between third-party e-commerce platforms, TikTok, WhatsApp, delivery apps, physical stores, and brand websites before making a purchase. The addition of AI has further compressed the product discovery path, with initial screening completed in a single conversation, meaning that the first recommendation a brand receives is likely to be beyond the direct control of retailers. 82% of Singaporean business leaders surveyed believe that large language models will become an essential tool for product discovery within the next year. Local retail professionals have begun adjusting their operational strategies, with 42% improving product content quality, 40% optimizing content structure for conversational queries, 39% submitting standardized data feeds to AI search platforms, and 38% rewriting product descriptions using natural language. The core goal is to enable machines to accurately identify and understand brand product catalogs. Traditional search only matches keywords, but AI assistants interpret user intent and provide direct recommendations for needs like products suitable for humid climates, running shoes for flat feet, or gifts under S$100. Brands with incomplete, inconsistent, or poorly structured product data will directly lose exposure in such conversational scenarios. Currently, only 28% of surveyed Singaporean enterprises have deployed agentic AI, and 52% of those that have not plan to launch it within the next six months. 86% of business leaders believe AI is raising consumer expectations, and 41% admit that meeting these expectations is now more challenging than ever. Implementing or expanding AI applications is both their top priority for the coming year and their biggest anticipated challenge. Agentic AI refers to AI systems that can take autonomous actions toward predefined goals, unlike ordinary chat tools that only generate text or answer questions. In commercial scenarios, this type of AI can help consumers compare product specifications, check inventory status, recommend bundled products, handle return requests, and transfer service inquiries. Fully automated procurement is still in its early stages, but AI's influence on product discovery and purchase decisions is rapidly increasing. Among companies in the Asia-Pacific region that have deployed agentic AI, only 5% are still in the use-case testing phase, while the largest share, 35%, is scaling applications across multiple departments and teams, including customer service, IT, and merchandise operations. Companies that have already implemented AI generally report improvements in customer satisfaction, personalization, operational efficiency, and employee productivity. These gains are particularly attractive in the Asia-Pacific market, where retail margins are thin and customer acquisition costs continue to rise. AI cannot automatically fix underlying system chaos; instead, it directly exposes the weaknesses of existing infrastructure. Among surveyed Singaporean business leaders, 81% say the number of suppliers their company works with has been increasing over the past two years, and only 25% have fully integrated customer data across sales, service, marketing, and e-commerce. Among companies with fragmented data, 36% experience slow or ineffective responses to customer issues, 32% cannot accurately measure the actual impact of business investments, and 40% bear high maintenance costs for dispersed systems. Omnichannel operations also show clear vulnerabilities, with only 2% of Singaporean multi-channel operators reporting no obvious failure points. The most common issue, cited by 46% of respondents, is the lack of real-time inventory synchronization, followed by inconsistent pricing and promotional information at 36%. These issues already exist in traditional channels and will be amplified with the spread of AI. If an AI assistant recommends out-of-stock items, quotes wrong prices, or offers advice that contradicts in-store staff, the customer experience breaks down. In markets like Southeast Asia, where consumers habitually compare prices across channels, such information mismatches directly lead to higher cart abandonment rates and erode consumer trust in brands. Physical retail is also being integrated into this digital chain. Survey data shows that 77% of consumers still prefer physical stores as their primary channel for holiday shopping, but the online connection to offline shopping is growing. 79% of consumers use their phones while shopping in stores, 12% consult AI assistants for purchase advice while standing at the shelf, and 86% of B2C respondents report that consumers expect the same level of personalization offline as online. Companies that have already made progress in data unification are seeing real returns, with 44% noting improved collaboration among sales, marketing, and e-commerce teams, 42% achieving higher customer retention and loyalty, and 31% experiencing better AI and automation outcomes. Data infrastructure has moved beyond backend maintenance to become a core factor directly impacting market competitiveness. As AI becomes the new shopping entry point, retailers need to ensure that product information, user context, promotional rules, inventory status, and service records flow seamlessly across all channels. Swatantra Kumar, Regional Vice President for Salesforce Asia-Pacific, stated that consumers no longer begin their shopping journey on brand websites or search bars; their starting point is an AI chat interface, social feeds, or delivery apps. Businesses that operate both online and offline need to reach consumers wherever they might discover products, and all of this hinges on having unified data that allows AI to accurately present their products and brand. For e-commerce professionals in Southeast Asia, the importance of AI is no longer up for debate; the real test is whether existing operational systems can adapt to a new consumption environment where the first brand touchpoint a consumer encounters is likely not a physical store or brand website, but a response generated by an AI assistant. This article was originally published on the Ebrun official website.