On 23 September 2026, SupplyChainTalk host Alastair Charatan was joined by Nanda Kishore, Head of Special Projects, Bergen Engines AS; Chris Wood, Former Director Exclusive Brands Logistics, Whole Foods Market; and Navneet Lekshminaryanan, Co-Founder and CEO, Holocene.
Views on news
UK importers of certain glass containers from China will face new provisional anti-dumping duties from 9 September 2026, following a decision by the UK Government to introduce measures after an investigation by the Trade Remedies Authority (TRA).
The new duties range from 24.65 per cent to 52.97 per cent, depending on the Chinese producer or exporter, and could significantly impact the landed cost of affected glass packaging imported into the UK. The new tariffs came with no warning, leaving no time for some companies to even take their stocks out of bond. How businesses manage immediate tariff changes depends on their agility and proactiveness, although in this case it was such short notice that they had practically no time to respond to changes to mitigate their impact.
Scanning international news in itself won’t make the business responsive and resilient; it needs data on a granular level to achieve that. In a broad sense, visibility should cover rules of origin, as well as the declarations suppliers make. As a general rule, it’s easier to manage unexpected tariffs if the contract with your supplier has provisions for which party should absorb tariff-related cost increases.
How can cycle times get shortened without risking stockouts?
To reduce cash tied up in stocks, the first step is to reduce variability. But even when a supplier’s lead time is short, you can’t minimise your inventory without a good understanding of the supplier’s business and collaboration with them, especially when operating in JIT. Segmenting the inventory by risk and value is also key to being able to reduce inventory while keeping risk low. The process starts with a forecast coming from the end user. Then the supplier must consult their suppliers to see if any bottlenecks are to be expected. Demand planning tools can only be effective if the business has a grasp of external factors too and full visibility of the business’s ecosystem. In other words, the execution part is just as important as the planning part in order to make inventories slimmer.
Even a well-functioning supply chain can’t be 100 per cent proactive – reactive works perfectly with safety stock, but a proactive approach is needed to respond well to disruptions. But connecting all the data sources – purchase orders, supplier milestones, shipments, inventory, production priorities – to get an end-to-end picture is a major step towards proactivity. Disruptions must get prioritised; not all of them are equally consequential. AI-driven decision support tools can further hone planning by identifying patterns and predicting disruptions. It won’t make the actual decisions for human experts, though.
AI is a perfect tool for sending reminders of a supplier’s previous delays when you’re placing a new order with them. Finding the balance between stockouts and overstocking is a balancing act, where the cost of stocks sitting in warehouses must be considered against the cost of disruption. With state-of-the-art tools, the quality of data has become less important than before as they can both augment data and self-improve. LLMs in their current form may not be the silver bullet for achieving end-to-end supply chain visibility. But you can apply AI to specific supply chain use cases, where you don’t need to analyse all available data, only what that particular use case requires. However, always make sure that you have the right experts to read the data that the AI has come up with and never feed past mistakes into models.
The panel’s advice
Winston House, 3rd Floor,
Units 306-309, 2-4 Dollis park,
London, N3 1HF
020 8349 4363
© 2026, Lyonsdown Limited. teiss® is a registered trademark of Lyonsdown Ltd. VAT registration number: 830519543