The five-year retail playbook: how to navigate the journey to true agentic merchandising

The long game: transformation that sticks

The smartest retail leaders know that true transformation rarely happens overnight. It’s not a simple software upgrade, but an organizational evolution.

That’s exactly what agentic merchandising requires; a multi-year shift in how everyday pricing and inventory decisions are made, governed, and trusted.

But while the journey may be long, the roadmap can be remarkably clear. Over the next five years, retailers can follow a deliberate maturity curve, from manual control to guided automation to autonomous optimization whilst protecting business rhythm and team culture.

Download our latest agentic merchandising white paper.

Why five years?

Three key reasons make that timeframe both realistic and powerful:

1. Capability building takes time. Teams need to learn new skills: data literacy, guardrail setting, scenario testing, and exception management.

2. Systems maturity is progressive. AI models need live data and feedback cycles before they can act reliably.

3. Cultural confidence must compound. Trust in machine-led execution grows through evidence, not promises.

Trying to force full autonomy in twelve months risks fear, confusion and, ultimately, rejection. Five years allows for adoption, adaptation, and ownership.

Year 1: get the foundation right before you build

The first year is all about instrumentation and alignment.

• Map pricing and inventory workflows: document every step (who decides, what data they use, and how outcomes are measured)

• Clean and connect data sources: Point of sale (POS), enterprise resource planning (ERP), inventory, and promotion data must flow together in real time

• Deploy assistive AI: start with systems that make recommendations (e.g., markdown suggestions, reorder proposals) without necessarily taking autonomous action

• Establish governance: create a small cross-functional group to define guardrails, approval logic, and success metrics

• Build literacy: train teams to interpret AI outputs, not fear them

Above all, focus on trust before transformation. This is the year to show how AI can make pricing and stock decisions more consistent, not more complicated.

Year 2: empower humans and AI to work together

Once teams understand and trust the data, the next phase is to let the system begin to act, to an extent.

• Define clear guardrails: set parameters for automated actions: margin thresholds, price ceilings/floors, replenishment limits, and promo caps

• Enable agentic execution in pilot zones: let the system make small, reversible decisions within those limits (e.g., 2–5% markdowns or daily replenishment orders)

• Monitor exceptions: humans review anomalies daily, adding nuance or overrides

• Measure impact rigorously: track time saved, decisions executed, and incremental profit or sell-through improvement

• Showcase wins: share stories internally to build trust (e.g., “the agent cleared $1.2m of aged stock with zero margin loss”)

Think of this almost like an internship for the AI. It’s learning the business rhythm, and the business is learning to manage an AI. Confidence starts to shift from individual instinct to institutional intelligence.

Year 3: collaboration and scale

By year three, trust and results have started to build momentum. This is when agentic thinking moves from pockets of excellence to systemic capability.

• Expand the scope: add adjacent processes: promotions, allocation, and assortment forecasting

• Unify data models: align elasticity, demand, and stock optimization models into one unified decision framework

• Introduce agent governance dashboards: senior leaders can now see every agentic action, exception, and performance metric in real time

• Integrate finance: open-to-buy and cash flow planning adjust dynamically as AI decisions flow through inventory and pricing

• Redefine KPIs: start tracking “decision latency,” “agentic accuracy,” and “margin protection,” not just sales and stock

By now, we’re firmly in the “collaborative era.” People and AI are operating in tandem, with the former setting the commercial intent and the latter executing it continuously.

Culturally, this phase feels liberating: meetings move from “what do we do?” to “how did it perform?” and “what do we tweak next?”.

Year 4: reorganizing and reshaping teams around intelligence

By now, the technology is firmly embedded, but the rest of the organization must catch up. This is the moment to align structure with capability.

• Flatten hierarchies: remove redundant approval layers in trading and merchandising

• Create new hybrid roles: e.g., AI trading partner, agent governance lead, commercial data translator

• Reskill planners and merchandisers: shift focus from manual reporting to designing agentic rules and scenarios

• Merge analytics and trading: create unified “Commercial Intelligence” teams that oversee both insight and execution

• Adjust incentives: reward decision quality and agility, not just revenue or sell-through

This is a crucial juncture where leadership courage really matters. If the structure stays static while the systems evolve, you end up with faster tech and slower people. The winners will be those who realign accountability, not just process maps.

Year 5: Retail that runs itself (within reason)

By the fifth year, agentic systems are executing the bulk of repetitive pricing and inventory decisions. People focus on strategy, supplier relationships, and innovation.

• Entrust end-to-end cycles: let agents run full markdown, replenishment, or promotion cycles with human review by exception only

• Automate feedback loops: AI learns from its own outcomes, continually improving elasticity and forecast accuracy

• Integrate multi-agent orchestration: pricing, inventory, and marketing agents coordinate actions automatically (e.g., price drops trigger email campaigns, low stock triggers reorder)

• Audit and refine governance: ensure explainability, fairness, and data security remain intact

• Measure total ROI: by now, impact should be visible across the profit and loss (P&L) statement—think higher margin, lower working capital, and faster decision velocity

You’ve moved from “AI-assisted” to “AI-accelerated.” Humans haven’t disappeared, but they’ve leveled up. The business operates continuously, not cyclically.

Aligning the C-suite to your agentic journey

For any transformation project of this magnitude, C-suite unity is essential. The CEO, chief financial officer (CFO), and chief merchandising officer (CMO) must all align around one core idea: agentic merchandising isn’t a tech project, but an operating model change.

The CEO sets the tone, framing AI as strategic, not experimental. The CFO ensures financial rigour and governance. The CMO and CFO translate outputs into customer value, like smarter pricing, better availability, and more relevant offers.

Meanwhile, the HR director leads capability building, ensuring staff see the agentic transformation as an exciting opportunity, not a threat to their role.

With every C-suite member singing from the same music sheet, they help to create a forward-thinking environment where agentic systems can thrive without friction.

How to communicate progress along the way

Transformation fatigue is real, especially for projects of this length and magnitude. Retailers can maintain momentum by communicating value in simple, commercial terms:

• “Our pricing agent now adjusts 12,000 SKUs daily within guardrails.”

• “Replenishment latency dropped from 48 hours to 6.”

• “We freed 2,000 hours of trading time last quarter.”

• “We improved full-price sell-through by 8% with no increase in discount depth.”

These statements translate AI achievement into real, tangible business impact, And that’s the exact type of language that keeps boards supportive and teams motivated.

Looking ahead: from AI adoption to AI orchestration

By the end of this journey, most retailers will have multiple agents working across their pricing, inventory, demand, and marketing processes.

The next horizon will be true agentic orchestration, where those systems coordinate autonomously, exchanging insights with each other to optimize total value, not just local metrics.

For example, a pricing AI agent might detect softening demand, signal marketing to trigger a promotion, and inform inventory teams to slow replenishment.

In another scenario, a supply chain agent could spot an inbound delay, prompting pricing to hold current margins in order to protect availability.

This is where agentic merchandising becomes a true competitive moat built around a retail business. Agile, integrated, and impossible to replicate manually.

Success is about structure, not luck

Retail has always rewarded those who can move fast and decide well. Agentic systems don’t replace that human instinct, but multiply it.

But success won’t go to the first to deploy agentic AI; it’ll go to the first to structure it properly, with clarity of ownership, aligned incentives, and transparent governance. Plus, with an organization design that’s built for continuous learning.

Five years from now, the most admired retailers will be those who mastered not just data, but decision design. Because in the agentic era, the competitive advantage won’t be knowing more. It’ll be deciding faster, smarter, and with unwavering confidence.

Are you at NRF 2026? Stop by booth #4040 to see agentic merchandising in action and talk to UiPath experts. We're also hosting an Executive Roundtable discussion with retail leaders and a networking dinner (along with our partner Roboyo) for current and future UiPath customers. Get more information and register for these events on our NRF page.

Not able to join us at NRF 2026 but interested in more info about agentic merchandising? Check out my other blog posts:

• Agentic merchandising: what it really means, and how it will transform retail pricing and inventory

• How to build confidence in your agentic inventory and pricing decisions

• Redesigning the retail operating model in the agentic era