Key Takeaways
* Organizations are seeking AI tools to monitor emerging situations like tariffs or geopolitical delays in real-time, which impact the supply chain, inventory, and pricing.
* AI solutions are needed for real-time operational decisions, such as dynamically rerouting deliveries to optimize efficiency and improve the accuracy of ETA during peak periods.
* AI agents use data insights, including metadata from customer-provided visuals, to generate tailored recommendations, fostering a more personal connection with customers.
The 2025 holiday season is upon us. It’s a stressful period for retailers, manufacturers, and logistics companies who must ensure that customers can order items seamlessly and receive their deliveries on time.
To navigate these demands, we recently convened experts to discuss how AI tools can help ensure a successful and profitable season. Our panel featured Yasser Sherrif, Cloud Architect with the IBM build engineering team, specializing in AI technologies and the Watson product suite; Itai Ben Shalom, Senior Director, Supply Chain and Procurement at Mark Jacobs (LMVH), who works closely on bringing the supply chain to a more mature place; and Michael Elkin, GigaSpaces CTO, focusing on AI-driven solutions.
The Expanding, Complex Holiday Spending Landscape
The nature of the season has changed significantly: the holiday season has become much more spread out. In the US calendar, it now begins as early as Labor Day. This extended period means that retailers, manufacturers, and logistics companies must remain vigilant throughout Q4 to meet customer expectations, posing significant operational challenges.
Ensuring that customers receive their gifts on time starts with the back end, specifically in procurement and the supply chain. Itai Ben Shalom highlighted that for brands like Mark Jacobs, the challenges often originate higher up in the supply chain, not just in the last mile.
How can you navigate geopolitical and regulatory hurdles?
This year has presented significant instability, impacting supply chain operations. Factors include changes in the U.S. administration leading to new tariffs, which can drastically affect the entire “food chain” of pricing and deliveries. The supply chain also has to contend with geopolitical issues like the Red Sea situation. These issues compound, raising the critical question for retailers: Do you have the inventory needed to service your customers?
Effectively, organizations need an AI-driven crystal ball to monitor emerging situations in real time and understand how they affect every link in the chain and organizations are actively seeking AI tools. Mark Jacobs, for instance, is exploring tools like Altana, which tracks the origin of items and can flag if a shipment might get caught by US customs. This heads up aims to avoid delays, which often force companies to switch from cheaper ocean freight to expensive air shipping, a critical concern during a challenging year for retail margins. Although the implementation of such tools might not affect the current season, robust AI usage in the supply chain is a definite focus for upcoming seasons.
Agentic AI: personalized experiences and conditional commerce
Beyond the back end, AI is fundamentally transforming the customer-facing experience, especially within the omnichannel environment where shoppers expect superb experiences across multiple places at once. Yasser Sharif noted that a significant investment is being made in AI agents for both B2B and B2C spaces. These agents do more than streamline transactions; they provide a personalized, real-time experience. For example, an agent can assess external factors, like the local weather at a shopper’s location, and then provide a relevant set of product recommendations.
In the near future, commerce will involve agents actively shopping on behalf of a business or a consumer. This enables complex conditional transactions, similar to setting a limit order in stock trading, where an agent might buy a product when a specific deal is available, or pay an invoice when payment terms offer a discount, tasks that are nearly impossible to manage manually.
However, this transition to Agentic AI brings new complexities, particularly around governance, security, and traceability. As agents are deployed, evaluation frameworks and observability are necessary, and payment providers are now investing in agentic commerce transactions to ensure that payments coming through an agent are tied securely to the consumer.
When shoppers want an integrated experience, such as ordering an item online for pick-up in a brick-and-mortar store, AI agents are critical. To address this, payment providers are investing in onboarding suppliers for merchants. When an agentic transaction occurs, the payment provider works to ensure the product comes from the correct supplier, matching criteria such as location, price, or specific payment terms. This requires seamless integration with ERP systems, such as SAP Ariba or Coopa. The AI platform matches the right supplier based on the consumer’s purchasing history or previous product insights, and then verifies if that supplier is approved by the payment provider before the transaction flows through.
Last-Mile Optimization: Real-Time Decisions Under Pressure
The last mile of delivery remains a major bottleneck, especially during peak holiday periods. Michael Elkin highlighted that large shipping providers, managing the chain from order to final delivery, face chronic challenges in effectively managing their network. The complexity is amplified because the last mile is often handled by a complicated network of third-party vendors.
These organizations are seeking AI solutions to be part of their core, ongoing operations, not just for forecasting or post-analysis, but for making real-time decisions. One of the most critical pain points is the accuracy of the Estimated Time of Arrival (ETA). Numerous parts of the chain can cause delays, and metrics on last-mile vendors that deliver on time on a typical day, can plummet during the holiday spike when they simply cannot handle the sheer volume. Michael gave an example of how a GigaSpaces customer in the Logistics space is using eRAG to improve margins by optimizing last mile delivery routes.
AI provides the solution to respond rapidly. Supply chain managers need immediate answers to questions like: “What will happen if we reroute a certain amount of deliveries from this last-mile provider to another one right now?”. AI can provide sophisticated, real-time responses to optimize efficiency and cost.
In contrast, organizations like Mark Jacobs currently manage their US last-mile delivery more simply by relying solely on UPS through a large partnership with LVMH. They prepare in advance by holding meetings with UPS, keeping the warehouse open longer, and clearly communicating “guarantee dates” on their website for delivery before Christmas. However, when contemplating expanding fulfillment to Europe, where they may need multiple last-mile providers, the need for advanced AI to manage complexity becomes immediately apparent.
How can you build trust through personalized connections?
In the current uncertain economic climate, consumers prioritize factors like quality and trust, seeking a more emotional connection with their vendors. The challenge is to maintain this personal connection while leveraging AI’s efficiency.
AI agents are central to providing this personalized experience. An orchestration platform uses data insights about shopping habits, locations, and typical purchases to generate tailored recommendations. Furthermore, AI is expanding the types of data used for personalization. For instance, a shopper can take a picture or video, and AI can determine the metadata from those visuals, using it to refine the shopper’s profile and suggest related items. This insight data is crucial for retailers to know their customers’ habits and refine their product catalogs based on purchases in specific regions or locations.
An exciting application connects regulatory requirements with customer experience. Itai Ben Shalom mentioned the EU’s Digital Product Passport mandate, which requires a scannable code on items for customers to track the supply chain regarding moral sourcing and sustainability. AI can leverage these required QR codes not just for compliance, but also to build a marketing tool that fosters a more personable connection with the customer.
How can you overcome the holiday crush with real-time AI?
When systems inevitably encounter unexpected situations, such as a sudden storm, a delayed shipment due to a geopolitical event, or an overwhelming holiday crush, AI must be ready to respond immediately. This is where real-time operational AI solutions become invaluable. eRAG from GigaSpaces is an AI solution that enables users to explore their enterprise data in real time using natural language conversations. Utilizing the familiar ChatGPT interface, the solution allows data exploration using real-time operational data. This gives customers a real-time handle on their business by simply asking questions, generating insights, and receiving suggestions on how to operate in emerging situations. This real-time response is proving crucial in organizations’ ability to keep a handle on key business metrics and respond to emerging situations resulting from unexpected tariffs or the geopolitical instability that has become part of everyday business dynamics.
The Black Friday Readiness Scenario
In this scenario, a retailer needs to know “Are we ready for the upcoming Black Friday?” concerning high-interest items like PlayStation and Xbox gaming consoles. eRAG assesses Black Friday readiness by integrating inventory data with external information, such as Google Trends data from the last three years. eRAG performs the following synthesis in real time:
- Trend Analysis: Fetches previous trends, showing, for example, a high and sudden interest (e.g., 50% week-over-week) in PlayStation and Xbox based on web searches.
- Inventory Assessment: It pulled core business data, including units on hand, required stock levels, and price.
- Risk Verdict: Combining these data points, eRAG generated an estimation of risk. For example, the system might determine that PlayStation 5 and Xbox are at a medium risk of running out of supply. The final readiness verdict might conclude that the momentum is very high for these products, and the company is only “partially ready” or “not ready”.
- Strategic Recommendations: Crucially, eRAG moves beyond mere analysis to generate strategic recommendations and follow-up actions on the fly. This is a blend of accessing data, analyzing insights, and suggesting potential solutions.
For instance, upon realizing an inventory shortage for PlayStation 5, the retailer can drill down and ask if reshuffling stock between different US regions might help. eRAG performs a detailed analysis of the geo-location of shops and stores, assessing which regions are under risk. Based on this, it generates a precise plan of transfer opportunities, recommending how many units can be shipped from low-risk areas like the Midwest to high-risk areas like the South. This plan of action, complete with exact units and amounts, provides a solution for an upcoming problem.
This immediate responsiveness can easily be applied to other use cases, such as monitoring order rates for an item and connecting that to external factors like weather. For instance, if an order rate spikes and a blizzard is forecast, the system could dynamically adjust the guaranteed delivery dates, offering a much more real-time operational response than previously possible.
The Power Behind the Interface: The Semantic Layer
The primary challenge for any organization is not merely building a connector to back-end systems like Salesforce, ERPs, or custom warehouse databases, the true investment is in understanding organizational knowledge.
This is achieved through the semantic layer. This layer is crucial for understanding all the complex processes, relationships, and the unique nuances and jargon within a business. When a supply chain manager asks, “Are we ready?” the system must know who is asking, what units are used for measurement, and which processes are encompassed by that specific definition of “readiness”. The semantic knowledge of the back end is what effectively powers these sophisticated, ad hoc interactions, allowing for seamless business machine operations.
Conclusion
The holiday season demands unprecedented agility. From procurement grappling with tariffs and geopolitical delays to retailers struggling to personalize the omnichannel journey, and logistics providers attempting to manage the complexity of the last mile, the need for real-time, data-driven decision-making is paramount. AI is shifting from being a theoretical tool to a core operational necessity. Solutions like eRAG, which combine familiar natural language interfaces with deep, semantic knowledge of enterprise data, assist organizations in surpassing the challenges of the holiday rush.
