Ask a supply chain director how AI will change planners’ jobs, and the discussion often starts with technology: which platform, which vendor, which pilot. Ask a planner, and the questions are different. Which decisions will the system make? Which exceptions will still need attention? And who will be accountable when a recommendation turns out to be wrong?
That difference matters, because the real impact of AI on planning will not come from better algorithms alone. It will come from changing how decisions are made, and therefore what planners spend their time doing. Advanced planning systems started this shift years ago; AI accelerates it by widening the range of decisions a system can handle. To see what that means, start with the job as it is done today.
The planner’s job today
Planning is still highly manual in most organisations, and part of that is unavoidable. When a supplier changes a lead time, a promotion has not been entered or a sales forecast reflects a commercial target rather than likely demand, planners have information the system does not; correcting the plan is a rational response to weaknesses elsewhere.
We see the same pattern with the majority of our clients, whatever their size or industry: when a serious disruption hits, teams spend a large share of their time firefighting, and manual allocation tends to favour the customer who complains loudest rather than the one where stock creates the most value.
Forecasting shows the consequences most clearly, because it is the best-researched part of the role. A recent study in the International Journal of Forecasting found planners manually adjusting around 75% of statistical forecasts. A 2025 follow-up covering more than 100,000 forecasts found those adjustments improved accuracy for only about half of the products, and that upward revisions in particular often made the forecast worse.
That does not mean human judgment should disappear; it is often simply pointed at the wrong place. A 2026 field experiment published in Management Science, run in a spare-parts business where demand is irregular and hard to predict from history alone, found that planners’ corrections to the information feeding the forecast, a missing promotion, an outdated lead time, raised profitability by close to 5%. Put next to the studies above, the pattern is simple: changing the final number rarely adds value, while fixing what the system cannot know often does. And that is exactly where AI starts to change the role.
From correcting plans to governing decisions
None of this is conceptually new. Planners have always prioritised their attention, and segmentation methods such as ABC/XYZ have long told them where to look first. What changes is the scale: instead of flagging where a human should intervene, systems can now take many routine decisions themselves, within limits the organisation defines. This is what we call the move from people-centric to decision-centric planning, and it changes the planner’s job in several practical ways.
-
Set the rules for automated decisions
Planners define thresholds, tolerances, service priorities and escalation rules. The question is no longer “what should the number be?” but “under what conditions can the system decide?”
-
Focus on meaningful exceptions
Attention goes to situations with real business impact: an unexpected supply constraint, conflicting demand signals, a trade-off between service, inventory and cost.
-
Evaluating Scenarios and Trade-offs
Rather than maintaining one plan, planners compare alternatives and make the trade-offs visible for the business.
Building alignment within the organization
A recommendation only creates value if the organization acts on it. Planners explain assumptions, highlight uncertainties, and help sales, finance, and operations reach consensus.
This also applies beyond demand planning: a supply planner whose system reschedules orders within agreed-upon limits can devote time to supplier risks, capacity conflicts, and working capital. In practice, however, the shift is proceeding more slowly than software vendors’ marketing would suggest.
Do companies need fewer planners?
The available evidence does not support a firm answer, and planning roles still appear on shortage lists in parts of Europe. More likely than disappearing teams is teams changing shape: transactional work shrinks or centralises, planners manage larger portfolios, hybrid roles emerge between planning, analytics and process governance, and the work that remains demands more judgment. One question cannot be skipped, though: experienced planners built their expertise on today’s processes and tools, and they will need a credible transition path, not an assumption that everyone becomes an analyst overnight.
Why AI progress often stalls
Because introducing AI is easy compared with redesigning the process around it. In a late-2025 survey, only 17% of supply chain leaders were doing that redesign; most were adding use cases one at a time. The familiar problems follow: outdated master data, commercial information that arrives late, unclear decision rights, and planners compensating in spreadsheets because the formal process does not reflect how decisions are really made. Adding AI on top of that environment does not remove those weaknesses. It can automate them.
What organisations need to do now
The starting question is not “where can we deploy AI?” but “which planning decisions are we ready to automate?”. Four actions follow.
-
Manage manual overrides
Define when planners can change a recommendation, when approval is needed, and measure whether interventions actually improve the outcome.
-
Develop decision-making skills
Understanding model errors, testing scenarios, quantifying trade-offs and communicating across functions matter more than tool training alone.
-
Gradually expand autonomy
Start with frequent, low-risk decisions, set clear escalation rules, and extend autonomy as performance is understood.
-
Redesign roles and decision rights
Be explicit about what the system recommends, what the planner validates and what management decides, and involve planners in that design: they know where the process works and where it breaks.
The real question isn’t whether planners will disappear
The biggest change AI brings is not a better forecast or a faster plan; it is that routine decisions can move away from planners, which changes where human judgment is needed. That brings us back to the director and the planner. The director asks which platform to invest in. The planner asks which decisions move to the system, which stay human and who owns the outcome. The organisations best prepared for AI will have answers to both.
A practical place to start is the oldest planning habit of all, the manual override: does anyone in your organisation measure whether changing the system’s recommendation actually improves the plan?
