Section 2
AI takes center stage, but infrastructure lags behind
Organizations are combining AI
and cloud technologies to
access better data and innovate
of organizations say adopting a generative-AI-first model will be key to getting cloud costs under control.
of organizations have reached a high level of generative AI maturity. Addressing human challenges and legacy technology issues will accelerate adoption.
of organizations strongly believe they have the right architecture and tools to enable data-driven decisions and support large volumes of data.
of organizations are investing a moderate to significant proportion (more than 6%) of their total IT spend in generative-AI-enabled data management.
Two-thirds of organizations believe they have already deployed generative AI to some degree, and the rest are in the early stages or actively experimenting (Figure 8).
European organizations are more likely than APAC organizations to be in the awareness or active stages of generative AI adoption. More than one-third (36%) of European respondents say this, compared with just 29% of APAC respondents. The U.S. sits in the middle: 33% say they are at these early stages of adoption.
But organizations must lay the foundation first to reap the full rewards. “With AI, people think it’s magic,” says Eurostar’s Laurent Bellan. “But if you want to have good AI, you first have to have good data. If you feed AI with poor data, you get poor results.”
Figure 8: Two-thirds of organizations believe they have adopted generative AI
Q: What level of generative AI maturity do you believe your organization is at now?
To make the most of agentic AI and its capabilities, organizations must first optimize their data storage and use. “We’re investing in infrastructure models and data models,” says Kellie Romack, chief digital information officer at ServiceNow. “Data is the fuel of AI, and part of my job is creating shared capabilities so we’re orchestrating across the enterprise rather than in silos.”
For IT executives, enhancing data capabilities is the most important way to release technology budget in the next 12 months. By feeding the right data into AI models and other technologies, organizations can make them more effective and efficient, freeing up time and budget.
Inefficient data capabilities also mean data costs are spiraling, and the majority (72%) of organizations view generative-AI-first models as a way to get cloud costs under control.
But despite this widespread awareness of the data imperative, only 36% of organizations strongly agree that they have the right architecture and tools to make data-driven decisions and support large volumes of data.
This figure (36%) remains unchanged since our 2023 study, despite the acceleration of generative AI over the last two years. Organizations must urgently address their data challenges to prepare for the next wave of innovation.
Nearly three-quarters (73%) of mainstream organizations are committing at least 6% of their IT spend to AI-optimized cloud infrastructure for generative AI. In Europe and the U.S., 72% of organizations in both regions are committing this level of spending. In APAC, this figure drops to 66%.
The high numbers in all regions mark a shift from simply using cloud services to relying on AI to create and manage cloud resources and IT operations.
When it comes to deploying generative AI, mainstream organizations say that the top two challenges are human ones:
Skills gaps and talent shortages
Resistance to AI-driven autonomy
Organizations must “bring humans along on the journey,” says Romack at ServiceNow: “We have to move AI and the technology that enables AI from a black box to a glass box. Fear is bred by lack of understanding. If we all understand more and have that democratized layer, it will help organizations to be more efficient and effective.”
For Innovation Leaders, the main generative AI challenges relate to infrastructure and data:
Testing, monitoring, tuning and debugging AI applications
Integrating with existing systems
Data silos and poor data quality / Lack of budget
For both Innovation Leaders and mainstream organizations, governance, security and privacy concerns are some of the least-cited challenges. Is this because organizations are taking care to align their businesses’ generative AI efforts with governance and security from the start? Or could they be underestimating the potential risks in this area?
Generative AI challenges vary widely across industries. While finance and healthcare organizations are grappling with trust and ROI, verticals like manufacturing, education, and government face deeper infrastructure and capability hurdles. The range of issues reinforces the need for industry-specific generative AI strategies.
Q. What are the top five generative AI challenges that your organization faces? (Number one challenge)
