記事を聴く
Key takeaways
- The constraint is changing. In the overall research sample, 87% of organizations run GenAI or Agentic AI in IT operations. At the same time, respondents expect an average failure rate of 43% for major AI projects initiated over the next 24 months. For retail and CPG leaders, this underscores the importance of organizational readiness alongside technological capability.
- Organizational barriers are prominent across the overall sample. Cross-functional coordination (40%) and managing organizational change and resistance (30%) are among the barriers respondents identify in pursuing their AI ambitions. These challenges are particularly relevant to retail and CPG environments, where decisions often span functions such as merchandising, supply chain, marketing and operations.
- The research also highlights the role of governance, workforce readiness and external expertise. Among organizations with very mature Responsible AI practices, 24% report difficulty balancing AI innovation with ethical considerations, compared with 41% of organizations whose approaches are not mature. Among respondents who used external expertise, 89% agreed that partners helped fill skill gaps and enabled progress.
Picture an inventory planner on Monday morning managing AI agents that continuously analyze demand, rebalance stock, test scenarios and identify the decisions that require human judgment.
Overnight, inventory has shifted across distribution centers and a promotion has been recalibrated using sales activity, weather patterns, social sentiment, supplier constraints and changing price elasticity. The planner begins her day by reviewing decisions already in motion, challenging the assumptions behind them and determining when human judgment should override machine confidence.
For retail and CPG, this illustrates how AI can change the work employees perform and the decisions they oversee.
Traditionally, planning often begins with gathering data, reconciling inputs and building a forecast. With intelligent systems increasingly able to sense, predict and act, employees can spend more time orchestrating those systems, interrogating their reasoning, managing exceptions and improving outcomes.
These capabilities are already being deployed across the organizations represented in our research sample. The findings indicate that AI is moving beyond supporting work toward performing a wider range of tasks.
Kristina Rogers, Chief Growth Officer, Retail, Consumer Packaged Goods and Luxury at HCLTech, puts the shift in perspective:
"Retail and consumer packaged goods companies have long been defined by the gap between business ambition and technology capability. Now that AI can sense, predict and act, competitive advantage will be determined by how quickly organizations adapt their operating models, workforce and decision-making to put it to work."
Retail's oldest complaint about technology just stopped being true
Retail and CPG organizations have experienced successive waves of enterprise technology. Enterprise resource planning (ERP), demand sensing, trade promotion and other platforms have often depended on lengthy implementation, integration and data-readiness efforts.
AI introduces a different set of possibilities and organizational questions.
Across the overall research sample, Agentic AI, systems that can plan, decide and act without waiting for a prompt, is used by 78% of IT organizations and 59% of production operations. Adoption across the sample includes:
- 87% of organizations run GenAI or Agentic AI in IT operations
- 86% in software development
- 79% in production operations
For retail and CPG, these adoption levels suggest that AI capabilities are becoming increasingly relevant to the systems and processes that underpin day-to-day operations.
Yet respondents in the overall sample expect an average failure rate of 43% for major AI projects initiated over the next 24 months.
The research suggests that technology capability is only part of the challenge. Organizational factors, including coordination, change management and skills, also shape AI progress.
Across the overall research sample, respondents identify several barriers to AI ambition:
- Cross-functional coordination: 40%
- Managing organizational change and resistance: 30%
- Shortage of AI expertise: 22%
For retail and CPG leaders, these findings reinforce the need to address the organizational environment around AI talent and technology.
“AI has moved from being a technology initiative to becoming an enterprise operating reality. What leaders are grappling with now is not whether AI can deliver value, but how organizations adapt their structures, decision rights and risk tolerance to keep pace with it.”
— Vijay Guntur, Chief Technology Officer and Head of Ecosystems, HCLTech
The implication for retail and CPG is that AI readiness extends beyond the technology itself to structures, decision rights, risk tolerance and workforce preparedness.
AI can require changes to the operating model and the technology stack.
Retail and CPG organizations have typically absorbed new technologies by implementing systems, training users and integrating tools into established processes.
Point-of-sale (POS), enterprise resource planning (ERP) and ecommerce platforms have all followed versions of this approach.
AI can create additional questions because systems may recommend, decide or act within business processes, requiring organizations to revisit accountability, oversight and decision rights.
When an agent reallocates inventory overnight, leaders need clarity on who owns the decision. When an assortment recommendation lacks context, such as regional dynamics, co-op commitments, or a delayed reset, employees need the literacy and authority to challenge the output and improve the system.
“Every AI initiative is, at its core, a change management initiative. The technology is rarely what fails. What fails is the human architecture around it - the behaviors, the trust, the willingness to work differently.”
— Jill Kouri, Global Chief Marketing Officer, HCLTech
These considerations provide a useful framework for thinking about different approaches to AI adoption in retail and CPG.
Two approaches to AI adoption in retail and CPG
The following “add-on” and “transformation” models are editorial frameworks for applying cross-industry research findings to retail and CPG; they are not segments identified by the study.
Across the overall sample, AI’s share of technology budgets is expected to increase by roughly 32% in the next 12 months, while the share of organizations devoting more than a fifth of technology spend to AI is expected to rise from 32% to 48%. The question for retail and CPG is how that investment translates into operational change and business value.
In an add-on approach, an organization may deploy new AI capability while leaving established processes largely unchanged. An AI forecast might feed the same monthly sales and operations planning meeting, while recommendations still move through processes designed around human speed. Change management and workforce preparedness may receive less attention than the technology deployment itself.
The overall research sample also points to tension between business and IT as AI adoption expands:
- 68% of respondents say business leaders are frustrated by IT's slow delivery of high-profile AI projects
- 62% say IT leaders worry about business teams advancing AI without oversight
Together, these findings illustrate why governance, coordination and operating-model design matter as AI scales.
In a transformation-oriented approach, organizations consider how decision rights, planning rhythms and workforce readiness may need to evolve alongside AI. For retail and CPG, this could mean aligning accountability with greater system autonomy, responding to signals more continuously and treating workforce readiness as a core part of implementation.
For this audience, the bar for that investment is the one Corrado Azzarita, Global Chief Information Officer of The Kraft Heinz Company, sets for everything:
“At the end of the day, only three metrics matter in business: revenue, cost and risk. If a provider cannot clearly articulate how a solution impacts at least one of those, then everything else is just noise.”
For retail and CPG leaders, this provides a practical business lens for evaluating transformation:
- Better-informed decisions can support revenue outcomes
- Stalled or poorly adopted deployments can add cost without delivering expected value
- Low workforce trust can introduce operational and governance risk
Three ways to apply the research to retail and CPG transformation
The following three moves are an editorial interpretation of the cross-industry findings for retail and CPG. They are not research-derived segments or practices attributed to a specific group of “transformation companies.”
Use governance to create clear permissions and accountability
Across the overall sample, 76% of organizations report delaying AI deployments due to Responsible AI concerns. For retail and CPG, clear governance can help employees understand what AI may decide, when human review is required and how exceptions should be handled.
The Responsible AI finding is specific: 24% of organizations with very mature Responsible AI practices report difficulty balancing AI innovation and ethical considerations, compared with 41% of organizations with less mature approaches.
Applied well, governance can give category managers and other employees clearer boundaries for AI use, overrides and escalation.
Build AI literacy alongside training
AI literacy extends beyond tool training to judgment: when to trust an output, when to challenge it, what a model may not see and how human context can improve the system.
The trade promotion model has never seen a supply shock. The demand model does not know the reset slipped. The person who knows those things is the most valuable component in the loop, but only if the organization treats her that way.
As AI capabilities evolve, retail and CPG organizations may need to make ongoing learning part of roles and operating rhythms rather than treat it as a one-time course.
Connect decisions across functions
In retail and CPG, AI agents can connect signals across merchandising, supply chain, marketing and store operations, for example, allowing a demand shift to inform inventory, media spend and shelf plans more quickly.
Achieving that coordination requires shared signals, clear roles and confidence in how decisions are made.
How the three AI impact imperatives can apply to retail and CPG
The AI Impact Imperatives 2026 identifies three imperatives across its overall private-sector research sample: the right foundation, the right governance and the right partnership. For retail and CPG, these provide a useful framework for considering workforce and operating-model readiness.
The urgency is underscored by respondents’ expectation of an average 43% failure rate for major AI projects initiated over the next 24 months. Applied to retail and CPG, the three imperatives can help leaders consider the data foundations, governance and external expertise needed to support AI adoption.
The right foundation: building trust
Retail and CPG employees need reliable data and context to assess AI outputs. When a demand model can incorporate signals such as point-of-sale data, retailer data and weather, planners have a stronger basis for evaluating and using its recommendations.
The right governance: creating clarity
Responsible AI can establish clearer boundaries for AI use and human oversight. Across the overall sample, 24% of organizations with very mature Responsible AI practices report difficulty balancing AI innovation and ethical considerations, compared with 41% of organizations with less mature approaches.
The right partners: supporting capability
Among respondents using external expertise, 89% agree that partners helped fill skill gaps and enabled progress, while 90% say partners accelerate time to value. For retail and CPG organizations, external expertise can complement internal capabilities as AI programs scale.
For retail and CPG, the question is how operating models, governance and workforce capabilities evolve alongside increasingly capable AI.
Return to the inventory planner at the start of the article. As AI systems take on more sensing, analysis and action, a practical leadership question emerges: if an agent makes or recommends a decision and an employee remains accountable, how should that employee’s role evolve?
An add-on approach may leave that question unresolved because roles and decision rights remain largely unchanged.
A transformation-oriented approach can redefine the role around governing outputs, challenging recommendations, applying business context and improving the system over time.
Retail and CPG organizations have spent decades expanding the range of what technology can support. As AI capabilities advance, the next phase is likely to depend on how effectively organizations align people, governance and operating models with those capabilities.
The full report provides the cross-industry research behind these findings, including:
- The ranking of organizational barriers to AI ambition across the overall research sample
- Respondents’ expected failure rates for major AI projects initiated over the next 24 months
- Technology-spending patterns and planned AI investment across the sample
- The criteria organizations use when selecting external partners
- Physical AI adoption data for supply chain and manufacturing contexts
The expected 43% average failure rate reflects respondents' expectations for major AI projects initiated over the next 24 months. It is not an observed failure rate or a direct measure of operating-model effectiveness. For retail and CPG leaders, it reinforces the importance of preparing the organization and deploying the technology.
Each imperative has a workforce dimension. Strong foundations can improve the information available to employees; governance can clarify decision rights and the acceptable use of AI; and external expertise can support skills and progress. Across the overall sample, 24% of organizations with very mature Responsible AI practices report difficulty balancing AI innovation and ethical considerations, compared with 41% of organizations with less mature approaches. Among respondents who used external expertise, 89% agreed that partners helped fill skill gaps and enabled progress.
