97% of Firms Can’t Prove AI ROI: 5 Expert Fixes

A staggering 97% of organizations currently struggle to demonstrate a clear return on investment from their generative AI initiatives, according to a recent survey of 600 data leaders. This widespread difficulty in proving business value presents a major hurdle for companies eager to capitalize on artificial intelligence’s potential. However, industry experts suggest several practical strategies can transform this challenge into a manageable process.

Begin with clear project boundaries and exit criteria. Gro Kamfjord, head of data at paint manufacturer Jotun, emphasizes the importance of having sufficient information to decide when to either scale an AI project or terminate it entirely. By modernizing their data infrastructure to a cloud-based system, her company created a centralized data hub that accelerates development. This foundational work allows teams to establish ballpark figures for expected business value. Kamfjord advises starting with simple, small-scale explorations. This approach provides the necessary insight to determine if a project will yield a payback, making the decision to stop a viable and often wise choice.

Securing organizational buy-in is critical for long-term success. Nick Millman, a senior managing director at Accenture, notes that CFOs rarely accept ROI calculations at face value. He recommends a three-part strategy. First, measure ROI using metrics that your specific business understands and values, avoiding overly complex or vague measurements. Second, ensure business stakeholders are fully aligned with the perceived value; projects lose credibility when only the data team champions the results. Finally, involve the finance department directly in building the business case. When someone from the finance team helps create the investment rationale, the CFO naturally develops a stronger vested interest in the project’s success.

Patience and persuasive storytelling are essential, especially when dealing with foundational investments. Boris van der Saag, an executive at Rabobank, points out that senior management often expects quick returns. He stresses the need to focus on long-term benefits and craft a compelling narrative for the boardroom. In his experience, a close working relationship between the data and finance functions transforms ROI discussions from one-way justifications into collaborative, two-way conversations. When the CFO begins asking how their team can change behaviors to unlock data opportunities, it signals a successful shift in the organizational dialogue.

Connect AI initiatives directly to overarching business transformation goals. Farhin Khan from AWS encourages leaders to pivot from mathematical ROI calculations to outcome-based impact assessments. She suggests delivering results in the language of the business stakeholder you are addressing. For instance, a Chief Marketing Officer will care about how AI personalization reduces customer churn. Furthermore, Khan advises explicitly linking each AI use case to the CEO’s strategic objectives, such as market expansion. Weaving this compelling narrative helps customize the message for different audiences and demonstrates how individual projects contribute to larger company ambitions.

Maintain rigorous oversight of all project components. Kenny Scott, a data governance consultant at EDF Power Solutions, highlights that effective ROI measurement depends on tight collaboration between IT teams, business stakeholders, and vendor partners. He warns against the “lone-wolf” tendency where individuals work in isolation. By building a modern data infrastructure, what he calls the engine room, his organization created a solid foundation for turning information into actionable insights. Scott insists that successful value delivery requires setting clear targets, outlining costs and expected returns, and adhering to deadlines. Understanding and controlling all the moving parts ensures the project remains on track and delivers tangible results.

(Source: ZDNET)