How agentic AI will change supply planning from firefighting to foresight

Being a planner in today’s supply chain operations can be a very challenging job, with each day potentially signalling a new shock to frameworks used just yesterday to make decisions.

As noted in the recent From Shock to Strategy: Building Value Chains for the Next 30 Years white paper by the World Economic Forum’s Global Future Council on Advanced Manufacturing and Supply Chains, volatility, uncertainty and ambiguity are no longer temporary challenges, but the new ‘normal’ that planning professionals and organizations across the entire value chain must master daily.

Managing the complexities of this new reality no longer allows for clinging to old practices. Instead, staying ahead demands a readiness to rethink established norms and embrace new ways of working.

One proven approach for organizations to increase their robustness is to move away from the performance of individual sites and shift to a holistic network perspective.

In such interconnected setups, plants – large and small – collaborate dynamically, pooling resources, capacities and even inventories to withstand shocks or capitalize on new opportunities. Network theory applied to supply chains underlines the value of combining large “hubs” with more flexible, smaller “nodes” to benefit from both scale and agility.

This move to network-based perspective, however, also multiplies complexity. Each production decision – what, where and when to manufacture – now spans a dynamic, multi-site environment.

Variables such as transit times, regional risks, capacity variations, local disruptions and sustainability demands turn planning into a truly multidimensional puzzle. In such a dynamic environment, it quickly becomes evident that even the best experts reach their limits when relying on spreadsheets and experience alone.

So how can planners master this complexity and volatility? Increasingly, they turn to intelligent planning tools powered by artificial intelligence (AI) and mathematical optimization.

These solutions integrate diverse data sources – from customer demand, inventories and factory capacities to constraints such as lead times, labour availability and geopolitical risks. Advanced algorithms process this data to generate scenario-based, optimized production plans across the entire network.

Such data-driven planning solutions compare well to a navigation system: suggesting optimal routes based on current traffic, the tool dynamically computes the best ways to meet demand and strategic goals across a global footprint – even as conditions shift daily.

Planners can also simulate disruptions – such as plant shutdowns, blocked shipping lanes, sudden demand drops – and immediately assess their impact. Consequently, they are now able to propose a multitude of possible responses within a very short time frame.

Importantly, these tools do not replace human decision-making – they empower it. By offering transparent, data-backed options, planners can focus on strategic objectives. They can quickly map out different scenarios according to various targets, such as demand fulfilment, cost efficiency or production stability.

This paradigm shift in planning not only changes how planners work – it fundamentally redefines their role: from operational coordinators to strategic orchestrators of conflicting stakeholder interests. The result is not only greater planning accuracy, but tangible benefits across multiple dimensions:

While intelligent planning tools already mark a major leap forward, the journey doesn’t end here. Instead, the next evolution in planning is already on the horizon – agentic AI.

While today’s data-driven tools already enable planners to navigate complexity with greater confidence, agentic AI promises to take this transformation even further. AI agents may soon autonomously monitor production signals, evaluate external events and even propose adjustments – all before a human even opens their planning tool.

These agents act as intelligent co-pilots: interpreting planning results, comparing them with real-time data from quality events, inventory thresholds, anomalies or large transactions. They can trigger alternative planning scenarios and suggest actions to planners. This shift allows for faster, more agile responses to disruptions – without sacrificing any accountability or feasibility.

With automation and AI increasingly embedded in planning processes, one question remains: Will human planners still be needed in the future? The answer is a resounding yes. In fact, their role becomes even more critical.

Faced with the challenge of coordinating a multidimensional and cross-functional system, planners are increasingly relying on agentic AI for tasks such as replenishment and rescheduling. However, especially in brownfield or less-automated environments, human-in-the-loop decision making remains essential.

Meanwhile, on the shop floor, constant flow with large series often faces trade-offs with short-term changes or process interruptions for flexibility. Therefore, the end-to-end responsibility for resolving these trade-offs still lies within cross-functional teams, although in a much more efficient way.

The planner – once fighting fires with limited insight – becomes the orchestrator of multiple, potentially conflicting stakeholder interests with a strong database at the core of his arguments.

Hence, organizations can reframe volatility in the supply chain from something to be feared to a source of competitive strength and innovation. In times of turbulence, such a transformation keeps humans and their responsibility in the loop.