Redesigning the retail operating model in the agentic era

It’s the elephant in the room across retail right now, and it rarely gets said out loud. But every boardroom is asking the same quiet question: how do we grow our business without growing headcount?

In a world of low growth, margin pressure, and high complexity, that question isn’t cynical but crucial to a retailer’s long-term survival.

Retailers are expected to deliver more decisions, more precision, and more agility with the same or smaller teams.

For years, automation has helped. Robotic process automation (RPA), better dashboards, helpful workflow tools. But these have now reached their limit. They make reporting faster, but don’t necessarily make decision making smarter.

That’s where agentic AI systems change the equation. They don’t just automate tasks, but automate judgment within guardrails. And that changes the shape of work entirely.

The changing anatomy of a retail business

Traditionally, retail organizations have been built around functional silos:

• Buying and merchandising: setting range and price

• Planning: managing OTB and replenishment

• Trading: optimizing sell-through

• Supply chain: executing orders and logistics

Each layer runs through multiple approval loops and trade meetings, with human gatekeepers overseeing insight and action.

Agentic merchandising compresses those loops, and when pricing, replenishment, and allocation can be executed automatically within agreed rules, the old “traffic control” structure starts to look outdated.

In the new model, teams shift from operators to orchestrators. Their job is not to push buttons, but to design and refine the rules under which intelligent systems operate.

A before-and-after snapshot

Markdown planning is a great example of how agentic AI reshapes a traditionally painful process. Today, merchandisers review sell-through and propose markdowns, planners manually model the impact on margin, finance checks everything against budget, and trading finally enters prices into the system. Five people often touch the same SKU before anything happens.

With agentic AI, the workflow becomes far more streamlined. The system generates recommendations automatically, surfacing only the exceptions for merchandisers to approve. Planners receive real-time projections instead of building models manually, finance monitors live margin dashboards, and the agent updates prices directly within guardrails. In the old model, action is slow and manual.

In the new one, the system acts instantly, with teams stepping in only when exceptions or strategy questions arise.

Why this doesn’t mean fewer retail people, but better people

It’s tempting to see this as headcount reduction, but the reality is subtler and, thankfully, healthier. Agentic systems reduce the volume of repetitive work, freeing teams to focus on higher-value activities:

• Scenario modeling and strategic planning

• Supplier negotiation and cost optimization

• Cross-functional collaboration between merchandising, supply chain, and marketing

• Interpreting signals and guiding the AI with new constraints or inputs

As manual noise falls away, commercial creativity can rise. Teams have more time to think about the bigger picture. The roles become more analytical, more forward-looking, and more commercially influential.

The best leaders will use this transformation to elevate their teams, not shrink them. Retail has always rewarded those who understand the customer and the numbers, and agentic systems amplify both skills.

A new retail operating model: from vertical to circular

In a traditional retail model, information flows up and down. From analysts, to managers, to directors, and back again. Decisions climb the hierarchy, get approved, then flow back to execution teams.

Agentic merchandising changes this to a circular model.

Here’s how it works in practice:

1. Agentic layer: executes repetitive actions (pricing, replenishment, allocation) continuously

2. Human oversight layer: handles exceptions and reviews weekly outcomes

3. Strategic design layer: refines rules, objectives, and commercial frameworks for the agents

4. Feedback loop: system performance data flows directly back to leadership dashboards

This circular loop is faster, flatter, and more transparent. Everyone can see how pricing or inventory actions are performing in real time, improving accountability across the board.

Rewriting the rhythm of retail with agentic AI

Every retailer lives by their calendar. Trade meetings, range sign-offs, markdown windows, open-to-buy (OTB) reviews. But those rhythms were built for an analogue world, where information moved slowly.

Agentic systems demand new cadences:

• Daily micro-decisions: agents adjust prices, inventory, and replenishment automatically

• Weekly oversight: humans review system outputs and update strategic parameters

• Monthly recalibration: leadership redefines objectives and guardrails

• Quarterly governance: audit, compliance, and financial reporting review outcomes

Instead of reacting to the past, teams start shaping the next cycle while it’s happening. This shift in tempo, from retrospective to real-time, separates the next-generation retailers from the rest and defines the winners of this exciting new era.

New roles require new skills

As this transformation unfolds, entirely new roles are emerging inside progressive retail organizations that blend commercial acumen with analytics.

The AI trading partner oversees agentic workflows across pricing and inventory, acting as the bridge between data science and commercial teams. The agent governance lead defines the rules, ethics, and approval pathways for autonomous decision making.

A commercial data translator turns merchandising intuition into codified logic the AI can act on, while an agent performance analyst monitors accuracy, exception rates, and business outcomes. These aren’t futuristic titles; several enterprise retailers are already hiring for them, signalling the new gold-standard skillset in modern retail.

The CFO’s perspective: redefining productivity

From a chief financial officer's (CFO’s) viewpoint, agentic systems answer the pressing question of increasing trading output without adding cost.

A replenishment planner managing 2,000 SKUs can suddenly oversee 20,000. A pricing team that used to review ten promotions a week can test 100.

The operating leverage this creates is significant, with more volume and more precision delivered by the same headcount. And because decisions are faster and more consistent, working capital efficiency improves. Inventory is held in smarter places, markdowns are targeted, and margin volatility flattens.

Common pitfalls to avoid

Transformation of this scale comes with risks. Here are some of the most common to keep in mind:

1. Assuming technology will fix a bad process. If you automate a bad process, you just get bad results faster. Redesign first, then automate when it’s ready.

2. Failing to define ownership. When AI agents start acting, who’s accountable for their decisions? Establish clear governance early.

3. Neglecting culture. Some teams feel threatened by automation. Frame it as empowerment, not replacement. The AI handles grunt work, freeing people for judgment and creativity.

4. Ignoring explainability. Leadership trust erodes quickly if they don’t understand why an AI agent acted a certain way. Make transparency a non-negotiable design principle.

A practical roadmap for organizational redesign

Retailers embarking on this journey should think in three phases:

Phase 1: Process instrumentation (0–12 months)

• Map out key pricing and inventory workflows

• Introduce assistive AI to augment existing teams

• Create a cross-functional governance group (merch, supply, finance)

Phase 2: Role evolution (12–30 months)

• Shift manual execution roles toward rule design and oversight

• Create hybrid roles; part merchant, part data translator

• Embed confidence KPIs into performance reviews

Phase 3: Structural Realignment (30–60 months)

• Flatten hierarchy around decision making

• Merge analytics and trading into unified “commercial intelligence” functions

• Integrate agentic performance reporting into board-level dashboards

By the fifth year, your business isn’t just “using AI.” It’s operating agentically, where pricing, replenishment, and inventory decisions flow seamlessly from strategy to execution.

How leadership should talk about the new retail operating model

Change fails when people believe it is purely about cutting jobs, and it succeeds when they understand it is about making their work matter more.

The message should focus on precision and empowerment rather than automation, showing how teams can spend more time on creative, high-value work, and less on repetitive tasks.

Leadership should make a point of sharing early wins such as hours saved, margin protected, and faster decision making to make the benefits feel tangible and real to their teams.

People follow clarity, and if you can paint a clear picture of this new operating model in action through real, day-to-day examples rather than abstract concepts, they will adapt much more quickly than you might first expect.

The future retail operating model: lean, intelligent, human

Agentic merchandising removes friction from retail processes, not people. The most successful retailers in the next five years will be leaner, but also sharper: fewer layers, faster decisions, higher trust between humans and systems.

Every business will have to choose: Will you redesign your operating model with agentic capability at the core, or bolt it on awkwardly around old legacy structures?

The retailers who get the people side right will extract ten times more value from the technology than those who don’t, and the winners will be the organizations that make that evolution feel not like a threat, but like progress.

Get your free copy of "The future of retail is agentic" white paper.