Where should supply chain leaders invest in 2026 to advance data-driven intelligence and/or AI?
2026 will see more organisations with complex supply chains investing to make their data more AI-ready. It has become apparent that many businesses’ plans to embed AI in their supply chain have faltered due to poor data quality and governance.
This is a serious obstacle to the transformation of supply chain decision-making and we can expect to see it stimulate investment in governance, quality and integration. AI is only as effective as the quality of the data it processes, and inaccurate or fragmented data will lead to flawed AI-driven outcomes. This in turn can have a negative impact on user adoption of new AI-driven systems or processes, creating a trust deficit.
Despite years of investment to improve end-to-end visibility, most supply chains still rely on fragmented insights and disconnected systems. This results in critical delays between capturing data and taking action based on the insights generated from it.
More organisations will invest in access to timely data as they realise it is the first step towards harnessing AI, enabling accurate forecasting and timely action. As more supply chain organisations recognise this, they will seek to improve data integration and preparation so they can embed AI to optimise their decision-making for the greatest possible ROI.
How will AI continue to reshape supply chain visibility and decision-making — and what will separate hype from real impact?
We can expect to see more buzz around generative AI and agentic AI in this context. Generative AI is one of the technologies that organisations can harness to produce decision-ready insights to accelerate the outcomes that businesses want. Agentic AI, operating more autonomously, can scan and analyse supplier data, shipping schedules and compliance updates in real time.
For perhaps the next five years, however, agentic AI will still need human supervision so that, for example, a manager can make the key decisions that reduce friction across operations. But to be effective, both generative and agentic AI require highquality, unified data.
Once they have AI-ready data, organisations can start using prescriptive insights. Instead of looking at what happened, they will be able to use AI to predict what will happen and recommend the best actions to take. Ultimately, the ability to measure the impact of these data-driven decisions is what will separate the hype from reality.
What are the risks or disruptions that are currently underestimated and how should data strategies evolve to address them?
The ongoing uncertainties in geopolitics are certain to continue to affect supply chains. Import and export controls can change rapidly as a result of sanctions and localisation rules, with consequences for procurement, inventory controls, compliance and global logistics operations.
There are already labour shortages in the transport and warehousing sectors which will be exacerbated by an ageing workforce in which retiring workers aren’t replaced by younger recruits. Supply chain organisations will feel the absence of such employees who understand the importance of real-time, optimised business decisions and have the necessary skills to execute them.
Threats from cyberattacks will also remain a serious concern in the logistics industry. Many risk programs track physical suppliers but ignore the underlying digital dependencies which makes organisations vulnerable. Attacks on these dependencies can halt operations across many partners.
Finally, we will continue to face disruptions from the effects of climate change and extreme weather. Organisations often prepare for single episodes of extreme weather through modelling, but they lack the ability to model correlated failures or ongoing trends that can combine to cause long delays. Organisations need to put this right, addressing infrastructure fragility and its ability to amplify supply chain shocks.
What is the most important step to take now, to ensure those investments create decision-ready data that teams can act on, rather than adding another disconnected layer of reporting?
Mark Holmes: Warehouses produce a huge amount of data that has little value without the ability to ingest, harmonise, analyse and act on in real-time.
If the definitions of events such as receipt, a pick completion, a production output, or a late dispatch vary by site or system, it is not possible to make successful use of AI. It leads down a path around consistent process flows which seems to be where the rest of content below is not going.
Clean, harmonised and unified data from internal systems, SKU attributes/naming, and external data sources/applications is necessary for AI solutions that assist in decision making, demand forecasting and predictive maintenance on machinery including AI-enabled robots and real-time drone inventory stock analysis.
The priority is to make data AI-ready by improving governance, quality and integration, to ensure that insights are timely and trusted. Organisations with strengthened master data, tightened governance and integrated operational systems have data that moves with minimal delay between all the relevant logistics partners and customers, the warehouse and planning and execution teams and the myriad of applications they interact with.
Organisations with harmonised free-flowing data will detect an exception quickly, understand its cause and enable corrective action in time to change the outcome with prescriptive ROI based decision making back to the LOB and directly into machines and warehousing applications as required. When organisations can measure improvements in response time and decision quality, they have a clear indicator that their data strategy is supporting real operational impact.