ROI Remains the Greatest AI Adoption Barrier

Business leaders and IT decision-makers confirm AI use will reach mainstream levels of adoption as organizations devote a greater slice of their IT budgets to AI implementations, according to new IDC research commissioned by Lenovo. The new global 2025 CIO Playbook highlights AI spending expectations by IT decision-makers globally to nearly triple in 2025 compared with last year. However critical challenges include uncertain financial returns on these investments and gaps in organizational readiness.

While most AI use cases have met business expectations, proving the return of these investments remains challenging—financial risk and uncertain ROI rank as the greatest barriers to AI adoption. This tension is magnified by a disconnect between growing AI investments and pervasive doubts among decision-makers about its value. Despite the forecasted surge in AI spending, business decision-makers are not unanimous in their optimism of its impact. 37% of management remain skeptical or have reservations toward AI, while approximately 9 out of 10 AI-adopting respondents said that AI has met their expectations. This highlights a significant divide between the unbound potential of AI and business confidence.

IT leaders expect AI to account for nearly 20% of tech budgets in 2025, driven by accelerated adoption of Generative AI use cases. While only 11% of enterprises are currently using GenAI-powered applications, this number is expected to increase almost fourfold to 42% in the coming year. IT operations, software development, and marketing departments are expected to see the highest level of GenAI applications.

“AI is a marathon and a sprint – requiring parallel efforts to move quickly to modernize systems while ensuring the future-proofing of tech stacks,” said Ken Wong, President of Solutions & Services Group at Lenovo. “Our research shows organizations need to simplify the design, deployment, and integration of AI solutions to demonstrate the impact of these investments. This will instill greater confidence and fuel future investments.”

The research also reveals several organizational readiness challenges. While ethical issues and biases in AI and machine learning were cited as the most complex or risky aspect of AI, more than half of global businesses do not have an AI Governance, Risk, and Compliance (GRC) policy in place. To realize compelling productivity gains promised with AI agents and assistants, organizations must also train and upskill staff, modernize IT systems to effectively integrate these tools, and establish organizational processes that help navigate the ethical and responsible use of these tools. 

The report also underscores the fundamental importance of data quality in delivering successful AI implementation. Ensuring data sovereignty and compliance, and availability of quality data were cited as the most important factors in successful implementations, whereas AI failures are caused most often by data quality issues, IT costs, and integrating AI with existing systems and processes. To this end, 33% of respondents said that their organizations will be developing data management capabilities in the next 12 months.

Despite the urgency to move AI agendas forward, enterprises recognize they can’t go it alone. Lack of skilled expertise is the most common reason for not investing in AI, while the research found that access to partners with strong AI capabilities remains one of the most important factors in successful AI implementation. “To harness AI’s transformative power, organizations need a data-driven strategy that ensures scalability, interoperability, and tangible business outcomes,” said Ashley Gorakhpurwalla, President of Infrastructure Solutions Group at Lenovo. “We believe a hybrid approach to AI is essential for delivering scalable solutions, driving measurable impact, and accelerating AI-powered business transformation.”