Globally, holiday digital sales in 2025 are projected to reach around US$1.25 trillion, with AI-assisted or agent-influenced orders expected to account for US$263 billion. Over half of Singaporeans (57%) plan to increase their holiday spending this year, and 56 percent are starting their shopping early, reflecting strong consumer confidence.
As spending once again increases, new consumer shopping habits come forward
One thing that has changed significantly in this year’s report is how consumers will be looking for products. The findings showed AI-powered shopping will boom this year, with AI traffic set to rise by 520 percent year-over-year and peaking in the 10 days leading up to Thanksgiving, but this will be more for research than actual purchases. In Singapore, the IAS’ 2025 Holiday Shopping Report found that 74 percent of Singaporeans are turning to AI tools for gift inspiration, with 27 percent using them to research product or gift ideas online. Top items for AI-driven research will include toys, electronics, jewellery, and personal care as thrifty consumers look around for ideas and value.
Retailers countering AI with AI
Retailers themselves are countering back with some AI-enabled help of their own to manage more traffic from AI sources, more orders from keen consumers, and more red tape from ever-changing trade policies. Deloitte has been surveying retail buyers to gauge their strategies as they navigate the run-up to the holiday season, concluding: “Newer forms of AI and advanced analytics, barely on the radar for retail buyers in 2020, may be helping to build resilience for the 2025 holiday rush.”
The report found that 78 percent of surveyed retail buyers leverage AI-enabled tools to enhance buying activities, while 74 percent specifically utilize AI to address challenges stemming from trade policy–related changes. Respondents who are using AI report improvements in several key areas, including:
- Better supply chain management. Using AI analytics to predict potential disruptions and optimize logistics
- Pricing optimization. Algorithms analyse market trends and consumer behavior to dynamically price items based on demand
- Product assortment optimization. AI solutions help streamline inventory management to ensure products are in the right place at the right time
- Demand forecasting. Predictive models anticipate customer demand to reduce overstock and stockouts
AI implementations are easier said than done
While these benefits sound like a great way to capitalize on tech-savvy consumers spending more this holiday season, implementing AI at scale for national or international retailers is not as simple as flicking on a switch.
Out of the box, foundational large language models (LLMs) lack the context they need to collect, analyse, and act on data. Retailers need a technology architecture to tap into and feed AI with contextual real-time data across their complex web of operations, which sprawls across suppliers, to warehouses, and in-store to online fulfilment, often on a global scale.
Something more is required for AI
This is especially important in the latest iteration of AI with the growth of Agentic AI applications. An AI agent doesn’t just follow pre-programmed instructions, but thinks on its feet, makes decisions, and adapts to new situations.
This requires real-time access to business-critical data and a contextual flow of information. That means, from an architectural standpoint, Agentic AI requires retail tech departments to rethink how to integrate solutions to maximize their potential.
Enter event-driven architecture (EDA), which provides a robust solution by decoupling agents from one another using an event broker, or a network of brokers called an event mesh. This more loosely coupled approach avoids rigid dependencies that make the systems brittle, hard to scale, and difficult to maintain – that plagued early microservices deployments, for example.
By using an event broker, agents can communicate asynchronously. This loose coupling enables the independent evolution of agents, allowing different teams to build and deploy their specialized agents without the need to coordinate complex dependencies.
AI success hinges on being event-driven for retail data movement
An event mesh, powered by agentic AI, is uniquely positioned to address the retail AI priorities for this holiday season. By creating a unified, real-time data network, an event mesh enhances supply chain visibility and agility, enabling quick responses to disruptions. The same system anticipates the AI-powered consumer shopper demands by continuously analysing data from various touchpoints, allowing retailers to adjust strategies rapidly. In so doing, it provides a foundation for true omnichannel capabilities.
AI-driven retail use cases
Let’s quickly look at five ways AI and an event mesh can supercharge retailers during this anticipated record holiday season. We aren’t talking hypothetically either; these five ways are here and now use cases I have been seeing across our retail customer base:
Out of stock? Not with intelligent inventory and demand prediction
AI-powered Inventory Management moves beyond basic tracking to predict demand and automate the restocking process in real-time. By integrating event data from IoT devices like shelf sensors and RFID tags, AI analyzes stock levels and predicts future needs, adjusting orders based on external factors like weather. This efficiency minimizes waste and prevents costly stockouts, significantly reducing labor and human error.
The smooth “silent shopping” experience
The “silent shopping” model uses AI to solve in-store operational issues discreetly, minimizing disruption for customers. Sensors noting sudden temperature fluctuations or in-store cameras detecting spills, for instance, generate events that are instantly routed to the relevant staff’s wearable devices. An AI within the system prioritizes these alerts based on urgency and staff location, ensuring quick, quiet incident management that boosts both customer satisfaction and staff efficiency.
The comfort of real-time loss prevention
Retailers are reinventing theft prevention by integrating multiple data streams for real-time risk analysis. Video feeds detecting suspicious behavior are correlated with point-of-sale and inventory data within an event-driven architecture. AI analyzes this combined stream to accurately identify potential theft events, immediately alerting security personnel with specific information, enabling a more discreet and effective approach to reducing losses.
Seamless omnichannel personalisation
Achieving “omnichannel excellence” means providing a unified, seamless customer journey across all touchpoints. AI uses an event backbone to capture interactions – from website clicks to in-store beacon pings – to build a continuously updated customer profile. This ensures that when a customer switches channels (e.g., moving from an abandoned online cart to a physical store), their context follows them, allowing sales associates to provide immediate, personalized recommendations.
Meet the AI store manager
The AI Store Manager acts as a central orchestrator for all store operations. It processes real-time events from foot traffic sensors, smart shelves, staff wearables, and even weather forecasts. Using this contextual data, the AI makes dynamic decisions: optimising staffing levels, adjusting store layouts, or initiating restocking. This continuous flow of information and instruction creates a highly responsive, efficient store environment, ushering in the future of retail management.
AI is not just for Christmas
Real retail success with AI hinges on the quality and speed of the data feeding those AI systems. For AI to deliver on the promises of optimized pricing, flawless supply chains, and personalized customer journeys, it requires a constant, contextual stream of data from every part of the enterprise – from warehouse sensors and supplier systems to in-store checkouts and online clicks.
This level of real-time, enterprise-wide data mobility is impossible to achieve with traditional, tightly coupled architectures. EDA and an event mesh, on the other hand, provide the required flexible, scalable, and decoupled foundation to transform raw data into the real-time context AI solutions need – to truly maximize sales and streamline operations during the crucial holiday season… and beyond.
Ush Shukla is a Distinguished Engineer at Solace. As an Enterprise Integration Architect, Ush has more than 13 years of experience leading diverse teams in the implementation of large-scale middleware solutions across varied business domains.
TNGlobal INSIDER publishes contributions relevant to entrepreneurship and innovation. You may submit your own original or published contributions subject to editorial discretion.
Beyond the HQ Hub: How Singapore can remain relevant in a world of job hugging