As artificial intelligence (AI) becomes a regular fixture in businesses worldwide, measuring its financial impact remains elusive. According to McKinsey's 2025 report, while 88% of surveyed companies use AI in at least one function, only about 40% attribute any significant enterprise-level earnings to it. This discrepancy has sparked a pressing question among business leaders: how should the return on investment (ROI) from AI be measured?

To address this issue, researchers Mika Ruokonen and Paavo Ritala propose three methods for calculating AI ROI. The first focuses on specific business functions such as processing time and error rates. The second coordinates measurements across departments, while the third evaluates AI investments as part of an enterprise-wide portfolio using financial indicators like net present value and internal rate of return. The key takeaway is that organizations should connect AI activity with verifiable business outcomes rather than solely relying on adoption figures.

However, the cost of implementing AI goes beyond software subscriptions and computing infrastructure. Businesses may need to invest in data preparation, system integration, employee training, process redesign, and ongoing human oversight, all of which can significantly impact the economics of implementation. According to Stanford Digital Economy Lab's 2026 report, these intangible costs accounted for 77% of the challenges faced in 51 enterprise deployments.

While productivity gains are evident with AI, they do not automatically translate into financial returns. For instance, a study published in the Quarterly Journal of Economics in 2025 found that an AI assistant increased customer support agent productivity by 15%, but this improvement did not necessarily result in equivalent cost savings for the company. Converting such gains into financial value requires detailed information about staffing, customer demand, service costs, and the expense of maintaining the AI system.

In conclusion, businesses must carefully consider the full cost of implementing AI and accurately measure its ROI to ensure it contributes positively to their bottom line. This involves connecting AI activity with verifiable business outcomes rather than merely focusing on adoption rates or time savings.