Follow ZDNET: Add us as a preferred source on Google.

ZDNET’s key takeaways

• Most business leaders struggle to prove the value of AI projects.
• Success comes down to storytelling, especially with the board.
• Focus on business outcomes and track your progress carefully.

Evidence suggests that many business leaders struggle to prove that an investment in generative AI delivers measurable returns.

More than 97% of organizations find it tough to demonstrate the business value of gen AI, according to a survey of 600 data leaders by Wakefield Research on behalf of technology specialist Informatica.

Also: OpenAI’s Altman calls AI sector ‘bubbly,’ but says we shouldn’t worry – here’s why

However, measuring AI ROI doesn’t have to be an intractable challenge. ZDNET attended a panel session and spoke with digital leaders at the recent Informatica World Tour event in London to discover five ways to measure the value of AI projects.

1. Know when to start and stop

Gro Kamfjord, head of data at paint manufacturer Jotun, said her explorations into AI suggest that business leaders must have enough information to know when a project should be stopped or pursued.

To boost growth across its regional offices, the company modernized its data infrastructure to the cloud via a partnership with Informatica and Snowflake. A new centralized data hub enables faster development, meaning teams can streamline their AI preparations.

“We’ve seen in this project that it is possible to create a ballpark figure of what you’re trying to achieve or at least point to the business value that will come from a project,” she said.

Also: Despite AI-related job loss fears, tech hiring holds steady – and here are the most in-demand skills

Kamfjord told ZDNET that business leaders who start their AI explorations with something simple and small can either scale up that initiative when the time is right or pull the plug entirely.

“I’m not sure that putting a number on the project is the most important thing,” she said. “What’s more important is that you get enough information to stop the project if you see that this project won’t produce a payback.”

2. Win hearts and minds

Nick Millman, senior managing director in the global data and AI team at Accenture, said judging the end-to-end value of AI projects is tough, and emerging technologies require an investment in data foundations that won’t deliver a short-term ROI.

“I’ve never met a CFO who just accepts whatever ROI calculation you put in front of them,” he said.

“Your success comes down to winning over the hearts and minds of the organization that AI is the right thing to invest in.”

Also: AI lifts some software stocks, leaves others behind – who’s winning and losing, and why

Millman encouraged digital leaders to take a three-pronged approach. First, measure ROI in terms that the business recognizes.

“I’ve seen so many different approaches, from mega spreadsheets that are tracking every single element through to vaguely measuring an increase in revenue. I don’t think there’s a right or wrong answer. But be pragmatic in terms of what works in your organization.”

Second, get the business involved: “Too many times it’s the data organization saying, ‘Here’s all the value we’ve produced.’ But you really need the business stakeholders to be fully aligned with that value. Otherwise, the project doesn’t maintain credibility.”

Third, ask the finance function for help: “You get someone who’s used to building business cases and ROIs, and then, by implication, the CFO has a more vested interest in the investment case for your project if someone on their team has helped create it.”

3. Foster two-way discussions

Boris van der Saag, EVP of data foundation at finance firm Rabobank, said organizations must be patient in terms of ROI if they’re going to invest in the foundational elements.

“You need to focus on the things you can eventually reap in terms of benefits,” he said, suggesting that business leaders should concentrate on the storytelling elements that emphasize the long-term goals of the investment.

“That’s important in terms of the conversation with the boardroom, because senior management is, by definition, less patient.”

Also: 4 ways to scale generative AI experiments into production services

In terms of his business, van der Saag reports to the CFO. The close working relationship between finance and data helps ensure that ROI isn’t just a one-way conversation but instead is a two-way discussion that enables new opportunities.

“Our CFO is asking our teams, ‘What can I do? How can I change my behavior? How can I change the behavior of my team to enable some of the opportunity that resides in the data?'” he said.

“If you get the storytelling right and you get people on that journey, you will see a change in the conversation, and it becomes much more of a two-way interaction rather than just selling individual use cases.”

4. Join the dots to bigger goals

Farhin Khan, UKI head of data and AI at AWS, is another business leader who encourages digital leaders to communicate the value of AI through storytelling.

“If you are communicating the outcomes of your project, you need to pivot away from the conventional thinking of what’s the ROI of your use case, from a mathematical perspective, to what’s the impact from an outcomes perspective,” she said.

“Deliver those results in the language of the business stakeholder you’re talking to. For example, a CMO will be interested in how an AI-powered personalization use case will help reduce customer lifecycle churn.”

Also: Consumers more likely to pay for ‘responsible’ AI tools, Deloitte survey says

Khan also encouraged digital leaders to connect the dots from their AI use cases to the business transformation being led by the CEO.

“If the business wants to expand into new markets, communicate how each of your use cases will contribute to the result,” she said.

“It’s all about weaving this compelling storytelling into your narrative that you can take back and customize to the stakeholder that you’re talking to.”

5. Track the moving parts of a project

Kenny Scott, data governance consultant at energy specialist EDF Power Solutions, said effective AI ROI measurement relies on a tight bond between the various parties involved in the project, whether that’s the IT team, business stakeholders, or vendor partners.

“You’ve always got to ask questions about the projects,” he said, suggesting that smart digital leaders will ensure everyone is acutely aware of their roles and responsibilities. “There can be a tendency for people to go lone-wolf and do something themselves.”

Scott has helped his organization build a modern data infrastructure, which he refers to as the engine room, including Informatica as the foundation, Snowflake as the core, and Power BI as the cockpit through which users turn information into insight.

He told ZDNET that successful value delivery is all about creating targets and managing expectations. Outline costs, expected returns, and stick to the deadlines.

“You need to be aware of the moving parts that are in there and ensure that they’re understood and controlled so that the project doesn’t run away.”

Get the morning’s top stories in your inbox each day with our Tech Today newsletter