A new study from SAP and Oxford Economics has found that companies around the world are increasingly seeing a positive return on investment (ROI) from AI, even as challenges continue to mount.
Infographic: SAP and Oxford Economics and the value of AI in 2026
While the amount the average global company spends on AI has increased slightly this year to $28 million, the ROI on that investment has skyrocketed. Globally, companies expect an ROI of 21% ($6.3 million) this year, up from 16% last year. This ROI is expected to increase to 38% ($15.9 million) in two years.
Agentic AI is central to these ROI expectations. Over the next two years, the average ROI of agent AI is expected to reach $17.6 million, more than four times last year’s estimates ($4.3 million).
These findings were revealed in the new global study, Value of AI Report 2026, which surveyed 2,600 executives in 13 countries.*
Sean Kask, Chief AI Strategy Officer at SAP, commented on the research: “AI has moved from experimentation to implementation, and that is starting to show real returns. But there is still a long way to go. Because AI without context – be it processes, data or governance – leads to activities without results at best and risks at worst.”
Table of Contents
AI is approaching enterprise maturity
While global investment in AI increased slightly from $26.7 million in 2025, there were significant changes in key markets. Investment increased significantly in Brazil, the UK, Australia and Germany, while leading markets such as China and India saw a decline in funding.
Today, almost a third of all tasks (30%) in the average company are powered by AI, a number that is expected to rise to 48% in two years. But while strategic investments in AI have almost doubled compared to last year to 17%, piecemeal approaches are still by far the most common (41%).
Some of this could be a leadership issue. Less than half of companies have a dedicated AI leader responsible for AI adoption (46%), a clear framework for AI development (52%), or even training on AI capabilities and risks (41%).
But despite these challenges, 69% of companies are satisfied with their current AI ROI, although more than two-thirds are not convinced that AI is achieving its full potential.
Part of this optimism can be attributed to agent AI, as over eight in ten (83%) companies say agent AI has moderate to very high potential to transform their business. Still, the technology is still in its infancy: Only three percent of companies say they are fully prepared for agent AI, while the majority say they are either partially or not at all prepared.
Global companies are tackling key AI challenges
Companies face a number of challenges in achieving ROI with AI, including data, people and governance issues.
Data quality remains the biggest challenge for global companies. The number of companies saying they are data ready for AI has decreased compared to last year, with 73% of companies disclosing issues with incomplete data. And that impacts day-to-day work, with 79% of companies experiencing rework, delays or backlogs due to poor quality AI output.
Similarly, companies are managing the impact of AI on the workforce. Almost eight in ten companies (78%) are either unsure or agree that their company’s upskilling is not keeping pace with the development of AI tools. And only one percent of respondents said AI will have no impact on their workforce planning. Meanwhile, the use of shadow AI is increasing every year, with 69% saying it happens at least occasionally.
“The next step in creating value will be to deeply integrate AI into contextual data and processes,” Kask said. “But companies around the world need to understand that AI often offers value that is harder to measure than expected and risks that evolve faster than most governance companies can keep up with. Companies are quickly discovering that AI governance plays a fundamental role in unlocking the value of AI.”
Governance is a key barrier to the value of enterprise AI. Only 12% of companies say either their capabilities or their processes and frameworks are fully prepared to effectively manage AI.
These problems could be exacerbated in an agent future. Today, 38% of organizations do not have a human-in-the-loop process for agent workflows, 37% do not have permissions and access controls for agents, and only 44% have a registry of agents in their organization. This is critical because more than two-thirds of companies (69%) either agree or disagree that they are deploying agents faster than they can manage them.
The future of AI is the autonomous enterprise
“It will not be easy to get real value from AI because it requires a new approach,” Kask concluded. “Organizations large and small must connect AI to the data and processes that drive their organizations and ensure it has the context and governance to deliver trusted results. This is what we call the autonomous enterprise. This is not a technical change, but a human one. Because only when agents, processes and people work together as a unit can you achieve real value.”
Value of AI: Research from SAP and Oxford Economics 2026
*Australia, Brazil, Canada, China, France, Germany, Italy, India, Japan, Singapore, Thailand, United Kingdom and United States.
https://news.sap.com/2026/07/business-value-ai-spiking-increased-adoption-agentic-expectations/
