Short answer: AI is taking over supply chain planning arithmetic and is nowhere near taking over supply chain management, and the employment data shows it clearly. The US Bureau of Labor Statistics projects logisticians to grow 18 percent between 2025 and 2035, adding about 44,900 positions to a 2025 base of 255,100 with 2025 median pay of $82,320. That is roughly five times the 3.5 percent projected across all employment and one of the strongest projections in the economy. Supply chains got more complex faster than they got automated, and complexity is what creates demand for people.
Forecasting Improved and the Problem Got Harder
Demand forecasting, inventory optimisation and route planning have been mathematical problems solved by software for decades. Better models improve those solutions meaningfully, and none of them addresses the thing that actually breaks supply chains.
The last several years demonstrated the point at scale. Pandemic demand shocks, port congestion, a canal blockage, semiconductor shortages, regional conflicts affecting shipping routes, and rapid changes in tariff and trade policy. Every one of these was a discontinuity, meaning an event with no useful precedent in the historical data a forecasting model learns from.
Models extrapolate from patterns. Discontinuities are the absence of a pattern. When one occurs, the forecast is not slightly wrong, it is irrelevant, and what determines the outcome is how quickly a person can find alternative capacity, renegotiate terms, reroute freight and decide which customers get short-shipped.
That is why supply chain roles grew through the most disrupted period in modern logistics rather than shrinking.
What Is Automated, and What Is Not
| Function | Status | What the professional does |
|---|---|---|
| Demand forecasting | Automated, improving | Judging when the forecast has stopped being valid |
| Inventory optimisation | Automated | Setting the service and risk targets it optimises against |
| Route and load planning | Automated | Exceptions, constraints and last-minute changes |
| Purchase order generation | Automated | Supplier relationships and terms |
| Track and trace | Automated | Acting on what it reveals |
| Warehouse operations | Increasingly robotic | Exception handling, layout, labour management |
| Supplier negotiation | Not automated | All of it |
| Crisis response | Not automated | All of it |
| Network design | Assisted by modelling | Strategy, risk appetite, capital decisions |
| Compliance and trade rules | Assisted | Interpretation and accountability |
Read the automated rows as a group. Every one is an optimisation performed against assumptions someone else set. Setting those assumptions, and recognising when they no longer hold, is the professional layer, and it has grown rather than shrunk.
The week a forecast becomes worthless
A supplier in a single region produces a component used across a product family. News arrives that the facility has stopped, and the restart date is unknown.
The planning system continues to produce orders against a forecast that assumes supply. Within a day, someone has to decide how to allocate remaining inventory between customers, which means choosing who to disappoint and being able to defend it commercially. Someone has to qualify an alternative supplier, which involves specification, sampling, quality approval and a price negotiated from a weak position. Someone has to decide whether to air freight at many times the cost, and that decision depends on contract penalties, customer relationships and how long the disruption is expected to last.
Meanwhile the sales team wants promises, finance wants the cost contained, and the operations team needs a production plan that will change again tomorrow.
None of this is optimisation. It is negotiation, prioritisation and judgement under uncertainty, executed at speed with incomplete information. Better systems make the underlying data visible faster, which is genuinely valuable and makes the human decisions better rather than unnecessary.
Why Better Optimisation Creates More Work, Not Less
There is a counterintuitive dynamic in this field that explains the projection better than any argument about capability.
When planning becomes cheaper and faster, organisations do not run the same supply chain with fewer people. They run a more complicated supply chain. They add product variants because the planning cost of each one fell. They promise faster delivery because the routing can support it. They open additional channels because the inventory can be allocated dynamically. They take on more suppliers in more countries because the coordination is now tractable.
Each of those decisions is rational, and together they produce a network with far more interactions than the one that existed before. More interactions mean more failure modes, more exceptions, more supplier relationships to manage and more decisions that fall outside what the system was configured for.
This is the same pattern that played out with spreadsheets in finance and with computer-aided design in engineering. The tool made the underlying work cheaper, the organisation responded by demanding more of it, and employment in the profession rose rather than fell.
It also explains why the roles that grew are the commercial and exception-handling ones rather than the planning ones. The arithmetic scaled. The judgement did not, because judgement is still performed one decision at a time by people who understand the business.
What to Know Before You Draw Conclusions
Warehouse automation is the real labour story. Robotics in fulfilment centres affects operative roles considerably more than any planning system affects professional ones. That is where the physical labour displacement in this sector is happening.
Complexity keeps increasing. More product variants, faster delivery expectations, more channels, more regulation, more trade restrictions and more sustainability reporting. Each addition creates work that did not exist.
Exposure measures are not employment forecasts. BLS published AI exposure categories with the 2025-35 projections and states plainly that exposure “does not imply job loss, productivity gains, automation probability, or wage effects.” Analytical roles score high on task overlap while logisticians are projected to grow 18 percent.
Resilience is now a board-level concern. Organisations that were caught out by disruption are investing in dual sourcing, nearshoring and scenario planning. All of that is professional supply chain work.
Data quality limits everything. Most supply chain systems run on incomplete master data, and no model fixes that. Someone has to.
Where the Work Is Growing
- Supply chain risk and resilience. Scenario planning, dual sourcing, supplier financial monitoring.
- Network design and nearshoring. Restructuring where things are made and held, which is a strategic and capital decision.
- Sustainability and compliance reporting. Emissions accounting, due diligence obligations and trade regulation, all expanding fast.
- Warehouse automation implementation. Someone specifies, integrates and troubleshoots these systems, and demand exceeds supply.
- Supplier development. Working with suppliers to improve capability rather than simply buying from them.
A Decision Framework for Supply Chain Professionals
- Planner running the system’s outputs. The most exposed position. The optimisation is automated, so move toward exception management, supplier relationships or network analysis within the year.
- Procurement or category manager. Strong position. Negotiation and supplier development are not automatable, and the analytical support available makes you more effective.
- Operations or warehouse manager. Your work is changing through robotics rather than through language models. Understanding automated systems well enough to specify and troubleshoot them is the growth skill.
- Entering the profession. This is one of the better-projected careers available. Enter through analytics and move toward the commercial and risk side, which is where the seniority is.
The test across all four: when the plan stops working, are you the person who fixes it? Supply chain plans stop working constantly, and that recurring failure is what sustains the profession.
It is worth adding one caution to all of this. An 18 percent projection describes the occupation, not any particular job inside it. Within a single organisation, a planning team can shrink while a risk and supplier development team grows, and both changes are consistent with the national number. For an individual, the relevant question is which side of that internal shift your current role sits on, and it is usually answerable by looking at whether your outputs are produced by a system that you operate or by decisions that you make.
Common mistakes right now
- Trusting a forecast during a period when its underlying assumptions have visibly broken.
- Optimising inventory without agreeing what service level the business actually wants to buy.
- Automating on top of poor master data, which produces confident and wrong outputs.
- Neglecting supplier relationships, which are the only thing that helps when capacity is scarce.
Building the Fluency the Growth Roles Require
The roles growing fastest in this field sit between the analytics and the commercial decision: people who can interrogate a planning model, explain why it is producing a strange result, and translate that into a decision a general manager can act on.
That requires understanding how these systems generate their outputs, where they degrade when conditions shift, what the data actually supports, and how to communicate uncertainty without either overstating or dismissing it. Learning that in a structured sequence is quicker than absorbing it from whichever planning platform your employer bought, and it transfers when you change employer or system. A certificate alongside operational experience makes it visible when a business is choosing who leads its planning transformation. If you want a structured route in, explore Coursiv AI lessons and check current plan details on the official site.
FAQ
Will AI take over supply chain management?
Which supply chain roles are most at risk?
Why did supply chain jobs grow through the disruption years?
Are autonomous trucks a threat to logistics employment?
Does warehouse robotics reduce total employment?
What should I learn to stay ahead?
Your Next Step
Take the last significant disruption your organisation handled and write down what actually resolved it. In almost every case the answer involves a person calling a supplier, making a prioritisation call, or finding capacity that no system knew existed. That list is the honest description of what this profession does, and building deliberately toward those capabilities is a better use of the next year than any anxiety about optimisation getting smarter.