Enterprise investment in artificial intelligence continues to grow, but many technology leaders are still looking for the productivity gains, savings and business value they expected. Recent reports point to a combination of infrastructure constraints, weak data foundations, workforce challenges and rising costs as barriers to realizing AI’s potential.
Executives’ confidence in returns has improved gradually this year, according to a report from SAP and Oxford Economics, but C-suite leaders remain unconvinced that their organizations are getting the full benefit from deployed systems. The problem is not simply whether a company can access AI models. Organizations also need suitable technology foundations, reliable information and a way to manage the operational and financial demands of adoption.
Infrastructure and readiness remain obstacles
A Kyndryl report identified external security shortcomings, aging technology stacks and workforce skills gaps among the factors limiting CIOs’ ability to benefit from AI. These issues can add cost and complexity to deployments, particularly when organizations must address existing technology limitations while meeting new AI demands.
AI readiness is also tied to the quality of information available to systems. A survey conducted by Harris Poll for Collibra found that more than half of enterprises are working to establish clear internal accountability for AI outputs. Separately, Dun & Bradstreet’s Gary Kotovets said the next phase of adoption will depend less on access to models and more on the quality of information those models can use.
That emphasis on data foundations has practical implications for spending. Organizations can invest in AI tools and infrastructure without ensuring that underlying information is suitable for the intended use. In that situation, additional technology outlays may not translate into stronger results, while the work of improving data quality and governance remains unresolved.
Budgets and AI billing are under pressure
Technology budgets are increasing as CIOs respond to AI demand, according to a West Monroe report. The report’s findings underscore a resource-allocation question for technology leaders: where to invest, and when to build tools internally rather than buy them. As spending expands, those decisions can shape both near-term costs and the long-term effort required to operate AI capabilities.
Consumption costs are another concern. A Splunk report found that more than half of businesses are failing to control AI costs amid growing interest in agentic AI. Billing spikes can place operational strain on organizations, making cost oversight an important part of deployment rather than a task to address after usage has scaled.
The challenge is especially relevant as enterprises experiment with systems that can perform tasks with varying degrees of autonomy. The available findings do not establish a single cause for rising bills, but they do point to a gap between expanding AI use and organizations’ ability to keep associated costs in check.
Change management is part of deployment
Technical readiness alone does not determine whether an enterprise AI initiative succeeds. Ismail Amla, senior vice president of Kyndryl Consult, said during a briefing with reporters that deployment requires change management across employees, systems and processes. He described the familiar categories of people, process and technology as more complex when human change is layered across each one.
That complexity is also visible in the adoption of AI agents. A report on the gap between executives and workers found that leaders are experiencing more success with agent deployments than employees, who face tool overload and trust gaps. Expanding access without addressing those concerns may leave organizations with growing usage but uneven adoption.
For CIOs, the reported obstacles are interconnected: infrastructure limitations and poor data can weaken outcomes, while skills and change-management gaps make it harder to embed tools in everyday work. Meanwhile, rising budgets and uncontrolled billing put pressure on leaders to connect spending with measurable value. The reports point to a broad readiness challenge rather than a single technical fix, with cost discipline, data quality and organizational preparation all affecting the returns enterprises can achieve.
