Cracking the Code: Proving Generative AI’s Business Value and ROI

Generative AI promises to revolutionize industries, yet for many organizations, translating that immense potential into tangible business value remains an elusive quest. Despite a projected global spending jump to $644 billion in 2025, a staggering 97% of businesses confess to struggling with demonstrating the measurable ROI of their Gen AI investments. This disconnect creates a significant hurdle for leaders facing pressure to justify burgeoning budgets.

The result is often a “proof-of-concept graveyard,” where promising pilots fail to scale and deliver real-world impact. But proving the worth of Generative AI isn’t an insurmountable challenge. The path to quantifiable returns lies in adopting strategic approaches to measurement and implementation, moving beyond the hype to practical, data-driven frameworks.

The Elusive ROI: Cracking the Generative AI Value Code

Generative AI holds immense promise, but for many business leaders, translating that potential into quantifiable returns feels like chasing a ghost. Evidence clearly suggests a widespread struggle to prove that investments in generative AI actually deliver measurable ROI. A recent survey, conducted by Wakefield Research on behalf of technology specialist Informatica, polled 600 data leaders and found that a staggering 97% of organizations are grappling with demonstrating the business value of Gen AI.

This isn’t just about skepticism; it’s a genuine challenge in a rapidly evolving tech landscape. As global Gen AI spending is projected to hit $644 billion in 2025, a 76% jump from 2024, the pressure is on chief data officers and other decision-makers to justify these hefty budgets with tangible results. Yet, a significant number of these initiatives are stalling, with two-thirds of surveyed data leaders admitting they haven’t successfully moved even half of their Gen AI pilots into full production. This creates a “PoC graveyard” where promising proofs-of-concept fail to scale into real business impact.

Beyond the Hype: Practical Approaches to AI Measurement

Despite the hurdles, proving AI’s value isn’t an impossible feat. ZDNET recently dove into discussions with digital leaders at the Informatica World Tour event in London, unearthing five key strategies for robustly measuring the value of AI projects.

• 1. Establish Outcomes and a “Stop Rule” with Finance: Before any Gen AI project even kicks off, you need to clearly define success in terms the business understands. Think revenue lift, churn reduction, or cost per contact. Gro Kamfjord, head of data at Norwegian paint manufacturer Jotun, emphasizes the need for enough information to know when to stop a project or push it further. This means creating “falsifiable hypotheses” with an agreed-upon “stopping rule” alongside your finance team. When finance is on board from the outset, the metrics, assumptions, and review processes are jointly owned, accelerating project growth.
• 2. Instrument Baselines and Run Real Experiments: Without a clear starting point, measuring progress is pure guesswork. Implementing processes to capture current performance *before* deploying AI is crucial. Think A/B tests or staggered rollouts to gather comparative data. Rigorous research design, for instance, has shown a 14% productivity increase for support agents using Gen AI assistants, highlighting the power of well-structured experiments.
• 3. Focus on Use Cases with Quick Paybacks: Not all AI projects are created equal in terms of immediate ROI. Prioritizing initiatives that offer quick, demonstrable value helps build momentum and prove the technology’s worth early on. Efficiency and productivity gains often provide the earliest tangible benefits, such as speeding up research, coding, or content creation tasks.
• 4. Model Cost Control and Total Cost of Ownership (TCO): Understanding the full financial picture is non-negotiable. This involves not just upfront investment but also ongoing operational costs, maintenance, and potential infrastructure upgrades. A clear handle on TCO ensures that the perceived benefits aren’t overshadowed by hidden expenses.
• 5. Turn Wins into a Scalable and Credible ROI Story: Successfully demonstrating value isn’t just about the numbers; it’s about effective communication. Digital leaders must translate technical successes into a compelling narrative for business stakeholders, especially the board. When AI is treated as a portfolio of products rather than isolated experiments, and successes are communicated clearly, the ROI story multiplies across different business lines.

Jotun’s Blueprint for AI Readiness

Paint manufacturer Jotun offers a compelling case study for navigating this landscape. The company embarked on a significant data modernization journey, shifting its data infrastructure to the cloud through a strategic partnership with Informatica and Snowflake. This move established a new, centralized data hub that significantly accelerated development times, boosting regional growth and streamlining their AI preparations. Gro Kamfjord noted that this project made it possible to create a “ballpark figure” for their objectives, allowing them to pinpoint the business value emerging from the initiative.

Jotun’s approach underscores the importance of a solid data foundation for AI success. The company migrated from on-premises PowerCenter to Informatica’s cloud-native, AI-powered Intelligent Data Management Cloud (IDMC) and from on-premises SQL to Snowflake. This combination delivers clean, quality, and governed data, critical for effective AI applications. By starting small and simple with AI explorations, businesses like Jotun can confidently scale up successful initiatives or, crucially, know when to pivot or pull the plug entirely.

As organizations push ahead with their Gen AI investments, the emphasis shifts from merely experimenting to rigorously proving tangible value. The insights from industry leaders and companies like Jotun provide a practical roadmap to turn the promise of AI into a verifiable business asset. Ignoring the ROI challenge simply isn’t an option.

Ultimately, unlocking the true business value of Generative AI requires a disciplined, proactive approach to measurement and strategy. By establishing clear outcomes with finance, rigorously instrumenting baselines, prioritizing use cases with rapid paybacks, and diligently modeling cost control, organizations can move beyond the “PoC graveyard” and transform their Gen AI investments into demonstrable, impactful returns. It’s about shifting from hope to hypothesis, ensuring every dollar spent contributes to a measurable bottom line.

Image source: Pexels