How AI is upgrading supply chain management
From predictive inventory to dynamic discounting, a raft of innovations are emerging
The World Bank says we are living through a “polycrisis”—a set of interconnected disruptions, from conflict to climate change to geoeconomic fragmentation.1 As these pressures compound, supply chain stress is reaching levels not seen since the height of the COVID-19 pandemic.2
In this environment, supply chain optimization becomes a board-level priority. Increasingly, leaders are turning to a new generation of single-interface working capital platforms, which help turn fragmented supply chain workflows into an integrated system—and use embedded AI to support faster decisions, stronger controls, and better risk mitigation.3
Here are three key areas of operations where AI is having a transformative impact.
Predictive inventory
Due to an increased focus on resilience, companies are shifting from just-in-time to just-in-case supply models.4 This involves stockpiling vital goods in easily accessible locations. The challenge is finding the sweet spot in inventory levels. “Having a buffer is great, but having too much inventory is a problem. Warehousing costs have gone up dramatically since 2020, while excess inventory ties up working capital,” says Dominic Giordani, Global Co-Head of Supply Chain Finance Product Management at J.P. Morgan Payments.
While traditional machine learning has long been used to predict required inventory levels, generative AI can improve decision-making by integrating huge volumes of external, unstructured data, including breaking news, trade policy, weather events, and social media. As a result, companies can optimize their just-in-case strategies more finely. Agentic AI could take this a stage further, by recognizing demand signals and automatically ordering more inventory to prevent outages or adjusting prices to clear excess merchandise.5 It could even ensure warehouse space is used with maximum efficiency, by tracking daily inventory movements and studying floor layouts.6 Across a warehouse network, it is estimated that AI unlocks an extra seven to 15 percent capacity.7
Multi-tier visibility
AI tools also allow for enhanced visibility across the supply chain. According to McKinsey, 95 percent of supply chain managers know their Tier One supplier risks, but only 42 percent know all of their Tier Two supplier risks, which can create compounding problems.8 Tier One suppliers are those companies an organization buys directly from, but Tier Two are the vendors that supply Tier One, so things can get opaque, especially when subcontracting is involved. “But companies have to know the full supply chain front-to-back because a severe disruption at a single supplier can go all the way up the chain,” says Giordani.
AI can help companies monitor and track complex, multi-tier supply chains.9 For example, it can be used to map suppliers by aggregating fragmented data on each vendor from customer records, invoices, financial statements, press releases, and news reports.10 This can also help get better insights into their operations and model risk. Once that map has been built, AI can craft and send communications at scale so that companies can keep in regular contact with all suppliers, helping surface any issues early.
Dynamic discounting 2.0
In times of stress, many larger companies push for extended payment terms to conserve working capital. But longer payment terms can shift that stress onto vendors and destabilize supply chains. An alternative is dynamic discounting: Instead of paying at net-30 or net-60 days, buyers offer suppliers early payment in exchange for a discount that varies depending on timing. Well-structured programs work for both parties, letting buyers reduce costs and suppliers smooth cash flow.
Dynamic discounting isn’t new, but AI is helping buyers better calibrate offers using factors such as payment history, supplier size, and balance-sheet strength.11 Generative AI tools can also scan emails and contracts to assess relationship context and generate personalized proposals.12 Tailoring like this was once impractical for multinationals with tens of thousands of suppliers; AI makes it feasible, improving acceptance.
Maximizing the discount rate isn’t always optimal for buyers. For example, it may make sense for a buyer to pay at net-10 for a larger discount; in other cases, the same buyer may prefer net-20 with a smaller discount to preserve working capital longer. These choices can shift with market conditions and business cycles. AI-based platforms synthesize internal and external data in real time to suggest discounting strategies.13
The biggest requisite for working capital platforms that seek to offer this kind of advanced AI functionality is reliable data. “All AI is built on data quality. Companies must break down silos, cleanse and standardize data, and enforce discipline,” Giordani says. Otherwise, AI recommendations may miss the mark. That’s why, for supply chain managers, the movement of data is becoming as important as the movement of goods.
https://www.jpmorgan.com/payments/payments-unbound/sources
Image credit: Lauren Joseph
