CEOs worry they are underinvesting in AI
A survey circulating through retail technology circles this month captures a tension that many leaders will recognize. Retail Dive reported on Cisco's 2026 CEO Survey, which polled 2,500 chief executives across 23 countries, and found that 69 percent view AI adoption as essential to remaining competitive. At the same time, almost two-thirds worry they are underinvesting in the technology. That gap between conviction and spending sits at the heart of the current moment in commerce technology. Leaders believe AI will decide competitive outcomes, and many fear their own commitment falls short of what the moment requires.
For commerce-tech leaders, the finding reframes the budget conversation. The question has moved past whether to invest in AI and toward whether current investment matches the strategic stakes. When two-thirds of chief executives express concern about underinvestment, the pressure flows directly to technology teams to demonstrate progress and to identify where additional spending would pay off. We read the survey as evidence that the mandate for AI has hardened at the top of the organization, which raises expectations for the roadmap that technology leaders present.
Infrastructure becomes the top constraint
The survey locates the primary constraint in infrastructure. More than half of the chief executives believe their existing infrastructure could limit AI initiatives, and 40 percent rank infrastructure modernization as their top business priority for 2026. Fewer than 25 percent said their networks are fully optimized for AI workloads. Those figures describe a foundation that many leaders consider unready for the ambitions placed on it. AI initiatives depend on compute, networking and data pipelines that were built for an earlier generation of applications, and the gap is now visible at the executive level.
This is a familiar problem in retail, where technology estates often carry decades of accumulated systems. Modernizing that foundation is slow and expensive work, and it competes for capital with customer-facing features that show returns sooner. The survey suggests chief executives now understand that AI ambitions rest on this groundwork, which gives technology leaders a stronger case for foundational investment. We advise framing infrastructure modernization as the enabling condition for every AI feature the business wants, because that is precisely how the C-suite is beginning to see it.
The data problem sits beneath the AI ambition
Beneath the infrastructure concern sits a sharper data problem. Only 19 percent of the surveyed executives said their enterprise in-house data is fully centralized and accessible for AI. That single figure explains much of the friction retailers encounter when they try to move AI projects from pilot to production. Models and agents need clean, connected and available data to function, and a large majority of organizations report that their data falls below that bar. The ambition to deploy AI collides with a fragmented data estate.
For commerce-tech leaders, this points to where the durable work lies. Centralizing and governing customer, operational and financial data is the unglamorous foundation that determines whether AI initiatives succeed. The 19 percent figure indicates how much room remains, and it aligns with the broader industry pattern of AI pilots that stall before scaling. We see data readiness as the gating factor for retail AI in 2026, and we expect the organizations that fix it to convert their AI investment into results while others remain stuck in experimentation.
Agentic operations need human oversight by design
The survey also offers a measured view of autonomy. Even as chief executives prioritize deploying AI agents alongside employees, 72 percent expect humans to retain oversight of AI systems through 2030. That expectation shapes how retail technology teams should design agentic systems. The near-term model places agents in support of human decisions, with clear checkpoints and accountability at defined points across commerce operations. Leaders are signaling that they want the productivity of agents together with the assurance of human review.
This has practical design consequences. Agentic workflows in retail, whether in merchandising, supply chain or customer service, need audit trails, escalation paths and controls that keep a person in the loop at defined points. Building these governance features from the start is easier than retrofitting them later. We advise commerce-tech leaders to treat oversight as a core requirement of any agentic deployment, because the executives funding these systems clearly expect it and because trust will determine how far autonomy can extend over time.
Partnerships and build-versus-buy decisions
The survey lands in an environment where partnership and build-versus-buy choices dominate the agenda. Industry analysts advise retailers to lean into technology partnerships to relieve operational pain in the short term while laying the rails for agentic operations over the longer term. That guidance reflects the reality that few retailers can build modern AI infrastructure and data platforms entirely on their own within the timeframe the market demands. Vendors and cloud providers offer a faster route to capability, and the ranked priorities in the survey reward speed.
For technology leaders, the decision is where to build differentiation and where to buy foundation. Customer data, loyalty records and the models tuned to a retailer's own demand are strong candidates to own and control. Underlying compute, networking and general-purpose AI services are strong candidates to source from partners. We advise mapping the AI roadmap against this distinction, so that scarce engineering effort concentrates on the assets that create competitive separation while partners carry the commodity layers. The survey's emphasis on speed and infrastructure supports that allocation.
What retail technology leaders should take from this
Three conclusions follow for retail technology leaders. First, the executive mandate for AI has hardened, and technology teams should expect scrutiny on both the pace and the return of AI investment. Second, infrastructure and data readiness are the binding constraints, with only 19 percent of executives reporting fully accessible data, so foundational work deserves priority and funding. Third, agentic systems should ship with human oversight built in, matching the expectation that 72 percent of chief executives hold through 2030.
The survey describes a leadership group convinced of AI's importance and uneasy about its own preparation. That combination creates an opening for commerce-tech leaders who can translate the ambition into a credible plan grounded in infrastructure, data and governance. The retailers that close the readiness gap will turn executive urgency into deployed capability. We expect the distance between conviction and execution to become the defining competitive variable in retail technology through the remainder of 2026.
