For decades, the supply chain discipline optimised around efficiency: lean inventories, just-in-time replenishment, and global sourcing calibrated to cost. That model created enormous value. It was also based on the assumption that disruption was a temporary state, and that operations would eventually return to normal.
In 2026, we’re all too aware that assumption no longer holds. Geopolitical volatility, trade realignments, regulatory pressure, climate events, labour shortages, and economic uncertainty increasingly overlap. Resilience cannot remain a contingency plan activated during a crisis. It must become part of everyday operations.
In response, industry leaders have often framed their response as a choice between just-in-time and just-in-case. Just-in-time reduces inventory and waste but can leave a business exposed when supply or transport is interrupted. Just-in-case adds safety stock and redundant suppliers, but can lock up working capital, increase carrying costs, and make it harder to respond when demand shifts.
Closing the gap between planning and execution
This approach presents a false dichotomy. Businesses should not have to make a permanent choice between efficiency and resilience. They need the flexibility to change their operating posture as conditions change – leaning toward efficiency when the environment is stable and adding protection when reliable signals indicate risk.
The new-era resilience model is no longer defined by how much inventory a company holds. It is defined by how quickly and reliably the company can understand a change, assess the trade-offs, and act.
Traditional supply chain management separates planning from execution. Teams develop a plan, release it, execute against it, and then measure what happened. However, in a continuously volatile environment, the assumptions behind that plan may change before execution is complete.
A demand shift, a port delay, or a supplier constraint can require immediate revision. Yet the signal often must pass through manual reports, sequential approvals, and disconnected systems. By the time the organisation responds, the most efficient options may have disappeared.
As such, planning and execution must increasingly coexist at the same point. That means continuously sensing changes in demand, supply, and logistics, evaluating their business impact, and coordinating action across functions. The delay between planning and execution becomes a structural vulnerability when the system is facing permanent disruption.
What an anticipatory supply chain looks like
Consider a beverage company operating across Southeast Asia. A change in temperature indicates that demand may rise in Vietnam, so an anticipatory system assesses available inventory, customer commitments, production capacity, and transport constraints. It identifies stock that can be repositioned from Singapore, prepares warehousing and logistics capacity in Malaysia, and signals the production requirements to the rest of the regional network.
The Line of Business owners have already set the objectives, business rules, risk tolerances, and escalation thresholds. The system acts within those boundaries, records the decisions it has made, actively scans for anomalies, and sends ambiguous or higher-risk exceptions to the designated person.
This is the shift from visibility to coordinated action. It depends on integrated planning and execution, trusted master data, real-time signals, and governed AI that understands business context. The goal is a supply chain that can respond before a disruption compounds.
The operating conditions differ across Asia Pacific, but the requirement for faster coordination is consistent. Chinese enterprises expanding internationally must manage more complex, multi-tier networks as they diversify production and serve new markets. Southeast Asia is becoming a more connected manufacturing and logistics hub, even as infrastructure maturity varies across countries. In Japan, labour constraints, industrial automation, and decarbonisation commitments make trusted automation increasingly important.
Building automation on context and governance
Enterprise automation cannot be added as a loose layer above fragmented process. AI must operate with the data models, permissions, process logic, and regulatory requirements that govern the business. This context is what allows an agent in life sciences to work within regulated manufacturing processes, or an agent in energy to support maintenance across distributed assets.
That’s where a solution like Joule can be used, providing an engagement and orchestration layer across SAP applications while purpose-built assistants and agents then support tasks in planning, procurement, manufacturing, logistics, and asset management. People direct the AI, the assistants orchestrate it, and the agents execute it. But at the end of the day, people remain accountable for strategy, objectives, and judgment, while high risk or exceptional decisions follow clear escalation paths to a defined human-in-the-loop.
Contrary to popular belief, governance does not slow this model down. When an organisation defines what a system can do, what requires approval, and what must be escalated, it can actually speed up processes while also providing control and a clear audit trail.
Global biopharmaceutical company Takeda offers one example. Through Autonomous Regulated Manufacturing, the company is achieving up to 10% productivity gains, up to a 25% reduction in revenue loss from stockouts, and up to a 5% reduction in safety stock. In the energy sector, SAP is working with RWE, one of Europe’s leading renewable energy operators, on an Autonomous Asset Management scenario where AI agents analyse thousands of past incidents, identify likely root causes, and prepare work orders with the right tools and proven fixes from other sites.
These examples connect automation to operational and financial performance while keeping people in control of consequential decisions.
Starting decision by decision
Supply chains will not become autonomous overnight. Leaders should begin with high-value, high-frequency decisions where delay creates measurable cost or risk, such as replenishment, order confirmation, and exception triage. They should connect the necessary data and workflows, define the boundaries for execution, and measure performance from the start.
South Korean electronics manufacturer LG Innotek reflects this directly. By partnering with SAP to build an intelligent supply chain analytics agent using SAP Integrated Business Planning and Business AI, they are liberating planners from manual data collection to focus entirely on strategic decision-making. This is a monumental shift toward a fully autonomous supply chain.
Across industries, the shift is the same. A leading Australian packaging manufacturer has taken the same approach. Through integrated business planning, over 30% of planner time has been freed from routine scheduling – redirected toward higher-value decisions and execution.
The companies that invest in this foundation do not just absorb disruption – they leapfrog their competition.
AI automation can then expand as results, trust, and organisational readiness improve. The businesses that win through disruption will be those with the data foundation, integration, and governance to act on what they know before disruption has time to compound.
Most importantly, organisations need to recognise that the measure of progress is not the number of AI pilots deployed. It is the business outcome: lower inventory, fewer stockouts, stronger service levels, greater productivity, faster response, and a better working capital position.
Abhishek Sahay is the Chief Revenue Officer for Supply Chain Management at SAP Asia Pacific.
