Enterprises willing to spend up to $250 million on gen AI, but ROI remains elusive

A sharp rise in enterprise investments in generative AI is poised to reshape business operations, with 68% of companies planning to invest between $50 million and $250 million over the next year, according to KPMG’s latest AI Quarterly Pulse Survey.

The report underscores a growing commitment to AI-driven innovation, with 67% of business leaders predicting that gen AI will transform their organizations by 2025.

“Our latest pulse survey confirms what we’re seeing with clients: organizations are doubling down on AI investments,” Steve Chase, vice chair of AI & Digital Innovation at KPMG said in a statement. “The data also shows growing momentum around AI agents, with over half of organizations exploring their use. Leaders are putting real dollars behind agents, but with mounting pressure to demonstrate ROI, getting the value story right is critical.”

Conducted among 100 C-suite leaders from organizations earning over $1 billion annually, the report highlights soaring investment, emerging challenges, and the evolving role of AI agents in enterprise transformation.

High expectations, but ROI challenges persist

Despite significant investments, only 31% of organizations expect to measure generative AI’s return on investment in the next six months. Notably, none of the surveyed leaders felt they’ve reached full maturity in their gen AI implementations.

The report suggested that the quality of organizational data remains a top obstacle, with 85% of respondents citing it as the most significant challenge for 2025. Privacy and cybersecurity concerns (71%) and employee adoption hurdles (46%) further compound implementation barriers.

“The dynamic nature of AI demands new ways to measure value — beyond the limits of a conventional business case,” Chase said. “As leaders work to define the right metrics, those measures must be tightly aligned with the business strategy and should account for the cost of not investing.”

2025: the tipping point for AI agents

The report identifies AI agents — autonomous tools capable of performing tasks and adapting in real-time — are emerging as key enablers for enterprise-scale AI adoption. Currently, 51% of organizations are exploring their potential to optimize administrative tasks (60%), customer service (54%), and business content creation (53%). However, only 12% have deployed such tools to date.

“Enterprise technology providers will introduce agentic AI capabilities throughout 2025, enabling organizations to move from experimentation and piloting to broad-scale deployment and integration into existing workstreams,” said Todd Lohr, Head of Ecosystems at KPMG’s US Advisory division. “The challenge will be scaling these solutions efficiently while maintaining oversight to enhance trust and foster human-AI collaboration.”

C-suite leads the charge, but adoption lags downstream

The survey revealed a significant disparity in generative AI usage between different levels within organizations.

Senior leadership, including C-suite executives, appears to be taking the lead, with 71% actively using gen AI tools. In contrast, only 26% of middle managers and a mere 15% of entry-level employees are leveraging these technologies.

This discrepancy points to a potential governance and usage gap within enterprises, where leadership adoption may not yet be translating into widespread operational use. Experts suggest this could hinder scaling efforts, especially as only 24% of employees are currently engaging with AI tools embedded into their workflows.

“Employee readiness remains a critical factor,” Chase emphasized. “Upskilling and seamless integration into workflows will drive adoption and ROI.”

Despite the challenges, there is optimism about driving greater adoption. Over 80% of organizations plan to integrate gen AI tools into formal performance development tracks, aiming to make these technologies an integral part of employee workflows and reduce barriers to usage. The need to align leadership’s enthusiasm with broader employee adoption remains crucial as organizations look to fully realize the transformative potential of AI.

“C-suite executives naturally have a distinct advantage in understanding gen AI’s potential to solve business problems, enhance workforce productivity, and provide a strategic edge,” said Abhishek Gupta, CIO at India’s leading satellite broadcaster Dish TV. “This explains the disparity in confidence levels between entry-level employees and executives. However, such disparity is normal and unlikely to negatively impact the success of Gen AI projects or the adoption of new initiatives.”

Shift in metrics reflects changing priorities

The KPMG report further added that for the first time since Q1 2024, productivity (79%) has overtaken profitability as the leading ROI metric for AI investments. Profitability remains a close second, surging from 35% in early 2024 to 73% by year’s end.

The report added that 54% of organizations are using gen AI productivity tools at least once a week and another 24% are using gen AI embedded into existing workflows at least once a week.

This shift underscores enterprises’ focus on leveraging AI for operational efficiencies amid macroeconomic pressures. According to KPMG, 88% of leaders continue to cite external factors as top influencers of AI strategy, underscoring the urgency of measurable results.

Data leadership, governance, and the road ahead

The KPMG survey stresses the importance of quality data, robust governance, and strategic oversight for scaling gen AI initiatives. Leaders prioritize scalability (66%) and technological expertise (61%) when selecting AI vendors, signaling a demand for enterprise-grade solutions that align with their business goals.

As gen AI adoption accelerates, enterprises face a pivotal moment: embrace AI’s potential to transform business or risk falling behind in a rapidly evolving digital economy. With multi-million-dollar investments and a focus on innovation, 2025 promises to be the year when generative AI turns ambition into actionable transformation.