Executive Summary
Retail is entering the second wave of AI deployment. The first wave (2022-2024) produced point-tool proliferation: demand forecasting here, personalisation engine there, fraud detection, inventory optimisation. Each tool worked in its domain. The problem that emerged by 2025 is that none of them learn from each other. At Walmart's Converge 2025 platform, the explicit design challenge was cross-agent coordination: not deploying more AI, but connecting the AI that was already deployed into a unified cognitive model of the retail operation. McKinsey's State of AI 2025 retail findings and Gartner's retail AI maturity surveys both confirm the sector-wide pattern: organizations with 10+ AI tools report lower overall satisfaction with AI program ROI than those with 3-5 tools in a coordinated architecture.
Retail's first AI wave left tools that don't learn from each other
Retail is entering the second wave of AI deployment. The first wave (2022-2024) produced point-tool proliferation: demand forecasting here, personalisation engine there, fraud detection, inventory optimisation. Each tool worked in its domain.
The problem that emerged by 2025 is that none of them learn from each other. At Walmart's Converge 2025 platform, the explicit design challenge was cross-agent coordination: not deploying more AI, but connecting the AI that was already deployed into a unified cognitive model of the retail operation. McKinsey's State of AI 2025 retail findings and Gartner's retail AI maturity surveys both confirm the sector-wide pattern: organizations with 10+ AI tools report lower overall satisfaction with AI program ROI than those with 3-5 tools in a coordinated architecture.
D2 advantage lives in the operating model, not in any single tool
D2: Digital Cognitive organization: applies to retail AI in a specific way: the competitive advantage in retail operations is not the intelligence any single AI tool has, but the intelligence the operating model has as a system. A demand forecast that does not inform the inventory optimisation AI is not part of a cognitive organization: it is a smarter spreadsheet. The Walmart Converge architecture is an attempt to build a D2 operating model: a system where data from one decision feeds the inputs of the next, and where the aggregate model of the retail operation learns from outcomes in a closed loop. Most retail AI programs are at an early stage of this transition. The D2 lens shows where the next investment should go: not more capability in individual tools, but more coordination between the tools already deployed.
The investment question shifts from "which capability?" to "how do we connect what we have?"
Retail transformation leaders with multiple AI tools in production face a portfolio coherence decision. The second wave of AI value in retail comes from the interactions between AI systems, not from the performance of any single one. This means the investment question has changed: it is no longer "which AI capability should we add?" but "how do we connect the AI capabilities we have so that the operating model learns as a whole?"
The Walmart Converge model suggests the answer requires platform-level design decisions: shared data contracts between AI systems, unified customer and inventory entity models, and a coordination layer that routes decisions across tools rather than running them independently.
Sector Context: Retail Has Moved Beyond the Point-Tool Phase
Retail AI grew function by function because the early business cases were easy to isolate. Forecasting could improve a demand signal. Personalization could optimize an offer. Fraud models could score transactions. Inventory tools could improve replenishment. Each investment had a local owner, a measurable use case, and a technology boundary.
That model now creates its own constraint. Retail is an interconnected system in which pricing changes demand, demand changes inventory, inventory changes fulfillment, fulfillment changes customer experience, and customer behavior changes the next forecast. When the AI systems governing those decisions operate independently, the retailer becomes locally intelligent but systemically inconsistent.
The second wave is therefore not primarily an AI adoption challenge. It is an operating-model redesign challenge: connecting signals, decisions, and outcomes so that the enterprise can learn as a whole.
Four Forces Driving the Shift from Agent Sprawl to Coordination
AI portfolios are becoming operationally interdependent. The more decisions delegated to specialized models and agents, the more frequently one system's output becomes another system's input. Unmanaged dependencies create conflicting actions and unclear accountability.
Customer journeys cross functional boundaries. A personalized promise is only valuable if inventory, pricing, fulfillment, service, and returns can honor it. Customer experience therefore becomes an orchestration problem rather than a channel problem.
Real-time retail compresses decision cycles. Pricing, availability, promotions, and fulfillment increasingly change within the same operating window. Batch coordination between functions is too slow when AI systems act continuously.
Measurement remains fragmented. Individual tools can report uplift while enterprise economics deteriorate through duplicated incentives, stock imbalances, service recovery, or margin leakage. Portfolio-level AI governance must reconcile local model metrics with enterprise outcomes.
The Structural Shift: From AI Portfolio to Cognitive Operating Model
A cognitive retail operating model connects the decision loops that already exist. Shared customer, product, inventory, order, and location entities create a common operational language. Data contracts define how systems exchange state. An orchestration layer determines which system has authority over which decision and how conflicts are resolved. Outcome data then flows back into the models and rules that shaped the decision.
D2 is the primary lens because the goal is organizational cognition rather than tool deployment. D3 provides the platform architecture that allows data and capabilities to be shared. D5 becomes important where associates, planners, merchandisers, and service teams work with AI recommendations. D4 governs the sequencing: retailers cannot replace the entire estate at once, so each investment must move the organization toward the target operating model.
Opportunities and Risks
Coordinated AI can improve availability, working capital, margin, service consistency, and response speed because decisions are optimized against shared context rather than isolated objectives. It also creates a path for new agentic capabilities without multiplying disconnected systems.
The principal risk is centralizing bad assumptions. A shared orchestration layer can propagate errors faster if entity models, permissions, or decision policies are weak. Retailers must also prevent automated optimization from creating unfair pricing, poor customer outcomes, or opaque decisions that frontline teams cannot explain.
Five Executive Priorities
Map the existing AI decision estate. Identify which systems make or recommend decisions, the data they consume, the outcomes they affect, and where their decisions collide.
Create shared retail entities and data contracts. Customer, product, inventory, order, location, and promotion definitions should not change from tool to tool.
Establish decision authority between agents and functions. Define which system can recommend, decide, override, or escalate for each high-value decision domain.
Measure portfolio coherence. Add enterprise metrics such as forecast-to-availability conversion, margin after service recovery, inventory productivity, and cross-channel consistency to local model KPIs.
Fund coordination before additional proliferation. New AI investment should demonstrate how it strengthens the shared operating model rather than adding another isolated capability.
Three signals that will show whether retail is consolidating or sprawling further
These signals: spanning platform performance data, industry maturity benchmarks, and emerging standards: will indicate which direction retail AI coordination is heading.
- Walmart Converge 2025 outcomes: public reporting on performance metrics from the Converge platform will set the reference standard for what cross-agent coordination in retail actually delivers: and what it costs to build.
- Gartner retail AI maturity benchmarks (2026 edition): will confirm whether the agent-sprawl problem is widening or whether the sector is genuinely consolidating toward coordinated architectures.
- McKinsey State of AI 2026 retail sector data: the 2025 retail findings identified the coherence problem; the 2026 edition will indicate whether retail organizations have begun addressing it structurally.
Executive Decision Test
The practical test for leaders is whether the next investment strengthens an enduring sector capability or merely improves one local initiative. Before approval, executives should be able to identify the operating-model dependency being changed, the reusable capability being created, the owner of the cross-functional decision, the outcome metric that will demonstrate value, and the governance mechanism that will remain after the implementation team leaves. If those answers are missing, the organization is still funding activity rather than redesign.
The sequencing principle is equally important. Leaders do not need to replace the entire estate before value can emerge. They do need each increment to move toward a coherent target architecture. That means using current initiatives to establish shared data, interfaces, decision rights, measurement, and reusable controls that subsequent initiatives can consume. The result should be cumulative: every deployment should make the next deployment easier, faster, safer, or cheaper.
This is also the distinction between adoption and capability. Adoption measures whether a technology or service is being used. Capability measures whether the organization can repeatedly produce the intended outcome under changing conditions. Industry Brief decisions should therefore be evaluated against capability compounding, not launch completion.
Closing Perspective
The central issue is structural rather than technological. The organizations that create durable advantage will be those that turn the capability described in this brief into part of the operating model, with clear ownership, reusable architecture, measurable outcomes, and governance that persists beyond an individual project. The leadership question is therefore not whether to adopt another tool or launch another initiative. It is whether the sector's operating architecture is being redesigned so that each investment strengthens the next one.
