The new skills of the demand planner of the future

The problem: the speed of change

Three facts explain why the demand planner is at the center of the biggest career transformation of the next decade:

• 40% of global jobs are exposed to significant changes due to AI (IMF, 2026)
• 9% of companies have reached true AI maturity (McKinsey)
• 66% faster is the rate at which required skills are changing for roles exposed to AI (PwC)

In other words: the gap between what the market needs and what the traditional planner delivers is growing, and fast.

‍

‍

Yesterday: the analyst planner

For two decades, the demand planner's job was reasonably stable:

• Manually extracting data from 5 different systems
• Building the forecast in Excel using moving averages and adjustments
• Reacting to stockouts that had already occurred
• Manually reconciling S&OP with sales and finance
• Reporting monthly accuracy, late and limited

It worked when volatility was the exception. Today, it is the rule.

‍

‍

Tomorrow: the orchestrator planner

The emerging profile operates on a different level:

• Agents prepare the data and features in real time
• ML assembled with confidence intervals replaces the moving average
• Demand sensing + external signals anticipate changes, they don't react to them
• Agents suggest actions; the planner decides and adjusts
• Continuous KPIs: accuracy, BIAS, margin impact

The difference isn't "doing the same thing faster." It's doing a job that is fundamentally different.

‍

‍

The 4 capabilities of the future planner

These aren't technical skills. They are working mindsets.

1️⃣ Data-fluent: knows how to read, interpret, and question model outputs. They don't trust blindly; they understand BIAS, identify when a model is out of its regime, and have the judgment to override it with a solid rationale.

2️⃣ Process-owner: designs what to delegate to agents and what to reserve for human judgment. They are the ones who decide the process architecture, not just those who execute each step.

3️⃣ Trade-off navigator: navigates supply chain tensions: cost vs. service, resilience vs. margin, speed vs. risk. They don't look for the right answer; they look for the right balance for the strategy.

4️⃣ Cross-functional: aligns sales, finance, and operations into a single plan. They translate the "truth of stock" into the language of sales, and the "truth of the market" into the language of operations.

‍

‍

Skills that are no longer optional

Behind these 4 capabilities lies a minimum technical stack that is becoming standard:

• Statistics: forecasting, accuracy, confidence intervals
• AI / ML: models, sensing, agents
• Tech: data, platform, integrations

You don't need to become a data scientist. You need to understand enough to engage as an equal with the data team and provide critical input to an AI agent.

‍

‍

The new process: Pyplan + AI Agents

The planning process is consolidating into 6 stages:

1. Ingestion / ETL: real-time connected data
2. Statistical forecast: baseline with classic models
3. ML sensing: captures external signals and adjusts the curve
4. Agent review: agents detect risks and propose actions
5. S&OP Consensus: cross-functional alignment with what-ifs
6. Continuous monitoring: real-time KPIs

The planner of the future doesn't execute every step. They design, supervise, and decide.

‍

‍

Three questions for your next career conversation

Before we wrap up this edition, three honest questions:

1. How many hours of your week are spent on tasks that a model could do in minutes?
2. Do you know how to read and challenge ML output, or do you accept the numbers because "the system said so"?
3. If your position disappeared tomorrow, what skills would make you immediately hirable?

‍

🚀 How Pyplan can help

The future of demand planning isn't less human. It's more strategic. Whoever develops these 4 capabilities and masters that stack will operate on another level, in added value, internal influence, and career potential.

Pyplan is an AI-native platform designed exactly for that model: AI agents that detect risks and recommend actions, with no black boxes and every node fully auditable. Companies like Nestlé and Pirelli are already operating at that standard.

‍

🔗 Discover Pyplan →

‍