McKinsey: State of Merchandising — Strategic Shift Toward Agentic AI

McKinsey & Company outlines a fundamental shift in merchandising. The function is moving away from a reactive, human-driven process toward a continuous, AI-powered system capable of acting in real time.

This is not just about efficiency. It reflects a deeper issue: the way merchandising operates today can no longer keep up with the speed and complexity of modern retail.

The conclusion is clear: The future of merchandising is real-time, autonomous, and execution-driven.

1. The Current State: High Effort, Low Responsiveness

Despite years of investment in analytics and digital tools, the reality remains challenging:

• Up to  40% of the time is still spent on manual work
• Decisions often come too late to make an impact
• 71% of organizations report limited impact from AI

Most merchandising processes are still:

• Periodic (weekly or monthly cycles)
• Fragmented (pricing, promotion, assortment disconnected)
• Lagging (based on past data)

At the same time, retail has become dynamic, store-specific, and real-time. A clear disconnect has emerged between how decisions are made and how the market behaves.

2. The Structural Bottleneck: The Execution Gap

For years, the focus has been on improving insights. But McKinsey highlights a critical shift: the main constraint is no longer insight. It is the ability to act on it in time.

In the current model, data is analyzed centrally, decisions are delayed, and execution happens downstream. By the time action reaches the shelf, conditions have already changed.

This delay is the primary source of value leakage in merchandising.

3. The Strategic Shift: Agentic Merchandising

This is where agentic AI changes the model.

Instead of linear workflows, merchandising becomes a continuous system that:

• monitors performance in real time
• detects issues as they occur
• determines the next best action
• increasingly… executes

All within a single loop: 

Not as a process but as an always-on system.

4. Quantified Impact

The potential impact is significant:

• Up to 40% of time can be reallocated to higher-value work
• Faster and more consistent decisions
• 0–3% margin uplift
• $3–5 trillion of commerce influenced by AI agents by 2030

This is not an incremental improvement. It represents a new operating infrastructure for retail.

5. Why Transformation Is Slower Than Expected

Despite the clear potential, many organizations are still early in the journey.

Key barriers include:

• Fragmented data across systems
• Legacy operating models built on periodic planning
• AI that remains at the recommendation layer
• Limited integration with execution in the field

As a result, many AI initiatives improve visibility, but do not materially change outcomes.

6. From Decision Support to Action

A critical distinction:

• Decision support systems → dashboards, recommendations
• Agentic systems → continuous decision + execution

The next phase requires: embedding decisions directly into workflows, enabling real-time action, and connecting decisions to execution.

Moving from understanding the shelf to acting at the shelf.

7. Implications for the Industry

This shift fundamentally changes how retail organizations operate:

• from planning cycles → continuous optimization
• from human coordination → system orchestration
• from delayed execution → real-time action

Success will depend not only on adopting AI, but on bringing decisions closer to the point of execution.

8. From Vision to Reality: What We See in Practice

McKinsey defines the shift clearly, and what is increasingly evident is the speed at which it is moving from concept to reality.

At effie.ai, we have operationalized an agentic approach to merchandising, establishing an instant closed loop at the shelf. The result is a step-change in merchandising KPIs and a twofold reduction in time spent by both merchandisers and supervisors.

In practice, adoption is refining the model itself. Real-world deployments are revealing consistent patterns and a set of emerging best practices that are shaping the future of merchandising execution:

1. Execution is already happening. AI is not just recommending actions; it is guiding and verifying execution directly at the shelf.
2. Decision-making is moving to the shelf. Actions are no longer centralized but executed in real time, in context.
3. The loop is becoming instant. Detect → decide → act → verify now happens within a single visit.
4. AI is evolving from analyst to operator. It is not only analyzing data but orchestrating execution.
5. This is capability transformation, not just efficiency. Beyond time savings, it enables consistent, high-quality execution at scale.
6. The execution gap is closing. The delay between insight and action is shrinking—and in some cases, disappearing entirely.
7. From awareness to intervention. Issues are not just identified but resolved immediately, in-store.
8. This is no longer theoretical. Agentic systems are already deployed and scaling across markets.

Full McKinsey perspective:

Link to the article