Enterprise AI has a participation problem disguised as progress. Adoption is everywhere, pilots are multiplying, and executives can point to no shortage of tools, teams, and internal experiments. But only about 6% of companies report measurable EBIT impact. Moving that number requires leadership to put AI inside the core business processes that shape financial performance, then redesign the work around it. The companies pulling ahead are making AI part of how the business actually runs..
Kjersten Moody, CEO of Elai North America, has served as the inaugural Chief Data Officer at both Prudential Financial and State Farm, and held global data leadership at Unilever overseeing a team of over 1,000. Across three Fortune 100 companies, she has delivered solutions with up to 30 to 40x ROI. She also co-founded High Peak, Prudential's first internal startup commercializing AI-driven lifespan prediction.
"If your AI work lives only in R&D or in corners of the business where underperformance won't touch the P&L, then overperformance won't touch it either. Moving the number takes the leadership, strategic courage, and operational precision to deploy AI in your core use cases," says Moody.
Moody identifies three patterns separating the companies that see real financial returns from the 94% that do not.
First, they do not treat AI as synonymous with generative AI. Their portfolios combine predictive and generative capabilities, deploying machine learning and statistical models alongside large language models. "It's really the predictive AI that is able to automate and improve decision making at scale into tangible value, also at scale. Predictive AI is going to have a resurgence in enterprise AI portfolios as companies want to join and expand the 6% club."
Second, they aim AI at core business processes, not peripheral experiments. If an AI project sits in a part of the business where underperformance carries no financial consequence, strong performance will not move the P&L either. The companies seeing EBIT impact have the operational discipline to put AI where the money actually flows.
Third, they rewire those processes end-to-end. "You're breaking out of this discrete chunk of a process and thinking about the true business beginning and the true business end," Moody says. That means articulating the future-state vision, rethinking how technology and people contribute to the redesigned workflow, and investing the capital to support comprehensive change rather than incremental bolt-ons.
Moody challenges the conventional sequencing that forces companies to fix their data infrastructure before starting any AI work. She argues that next-generation platforms can automate data engineering, model creation, training, testing, deployment, and observability in parallel rather than in sequence.
"Leaders need to do their homework on what this next generation of companies and capabilities represents. Be very careful of bias in strategic thinking that says you have to do your data and then you can do your AI. There are companies rapidly changing that paradigm." The shift matters because it frees capital for the broader process redesign and workforce investment that end-to-end rewiring demands, rather than consuming the budget on multi-year data modernization programs that delay AI value indefinitely.
The workforce conversation, Moody argues, is using the wrong unit of analysis. Companies keep asking what happens to jobs; instead, companies should start by asking how people will spend their time. "A job is a grouping of skills for how people are going to use their time to deliver business needs. If you start to think about time and skills, then you can repackage it into what that job of the future looks like and bring people into that new way." She says the fear factor comes from leaders using old speech patterns that frame AI as a threat to employment rather than a reallocation of how people spend their working hours.
At the board level, Moody sees a governance gap. Directors need to ask management where AI is required for the business strategy to succeed, not just where it is being used. "If the answer is that success does not require AI, then the incentive structure is now making AI optional versus required."
The impatience around AI ROI reflects something real. Unlike cloud or virtualization, AI is a technical revolution that affects the entire business system in many ways. "There's a race to be in that first generation harvesting value from it, to not become competitively obsolete. The more thoughtful, the more empathetic companies are going to be the winners in the AI age."
