The Real Barrier to AI ROI: People, and Their Resistance to Change

AI isn’t failing to deliver ROI. The problem is leadership.

Every headline that piles on about AI’s “lack of ROI” feeds the idea that the technology has overpromised and underdelivered. The narrative is simple and sticky—but it's wrong. AI is delivering better ways of working faster than most organizations and people are able to adapt, and much faster than the outdated ways we measure return on investment.

What’s stalling value isn’t the tech; it’s leaders chasing shortcuts: another product to buy, a quick fix, another proof-of-concept to show the board.

Like every meaningful transformation, the only path to value is consistency and commitment—and the same is true for AI. Move projects from pilot to production, embed AI into real workflows, retrain teams, track adoption and outcomes, and keep iterating. Think about this: What was the ROI of the Internet in 1995? Was that measurable at the time, knowing what it has fueled by 2025? 

ROI grows with disciplined execution, willingness to change, and patience. Frustration with the lack of traditional ROI doesn’t mean value isn’t there — it means we need to evolve how we adopt and define it.

Why ROI Looks Elusive

Traditional ROI models were designed for linear technology bets: you buy a system, deploy it, and count cost savings or efficiency gains. AI doesn’t fit neatly into that framework. Its value comes in speed, scale, and compounding improvements over time.

This point was underscored at the Wall Street Journal’s Tech Council Summit recently. CIOs agreed that traditional ROI metrics “miss the point of AI.” Instead of measuring lines of code written or minutes saved, executives argued that the real value comes from innovation and long-term transformation. As one leader put it, companies should identify high-impact projects and build the infrastructure to scale them rather than looking for short-term productivity spikes.

That perspective matters because it reframes the ROI debate. Of course you won’t find instant, P&L-moving impact from AI pilots. That doesn’t mean AI is failing. It means we’re looking in the wrong places.

What’s Really Missing

When news reports say “AI is failing to—deliver”—here’s another one from the Financial Times—it’s almost never about the technology. It’s about how it’s used and measured. Too often, companies:

• Use AI as an add-on instead of rethinking processes with AI as the foundation.
• Measure it narrowly, focusing too heavily on cost takeout instead of speed, scale, and adaptability.
• Avoid the harder work of changing how people work, relying on patches instead of transformation.

As I’ve said to clients, when AI “fails,” it’s almost always a skills issue, not a tech issue. The technology already outpaces most people’s ability to use it. Every stumble is not proof that AI doesn’t work; it’s a sign that leaders haven’t created the conditions for adoption.

And we don’t need to speak in hypotheticals about where ROI might come from someday. It’s already here.

For one client, a manual reporting process was modernized into an AI-powered platform that scaled with their growth. The change eliminated inefficiencies, improved accuracy, and enabled faster compliance reporting. The result: average annual savings of $30,000 per large customer and $2,000 per smaller customer—measurable, repeatable, and real ROI today. For another, AI compressed a project timeline that used to take months into just three weeks, while boosting output by 50% and doubling developer productivity. That’s not “potential.” That’s value on the table right now.

These examples may look like ROI in the more traditional sense—and they are. But that's OK. You can still calculate ROI from AI using today's tools and math while recognizing that additional ROI will come. It exists in scalability and compounding benefits that can be harder to track or remain “hidden” until organizations evolve how they measure it. If you can’t measure it perfectly today, that doesn’t mean ROI isn’t there—it means you should revisit how you calculate it.

Redefining ROI for the AI Era

The truth is simple: ROI in the age of AI cannot be defined the same way it was for ERP, cloud, or automation. The real returns show up in dimensions that don’t fit neatly into quarterly financials:

• Speed: How quickly you can move from insight to execution
• Scale: How much more you can produce per individual human
• Compounding value: How AI-driven processes learn, improve, and deliver greater returns the longer they’re in use

That’s why the question isn’t “Where is the ROI?” It’s “Are you measuring the right things?”

When leaders continue to define ROI solely in terms of near-term cost savings, they miss the bigger payoff. They undervalue the speed advantage of faster decision-making. They ignore the ability to scale output without scaling headcount. And they discount the compounding effects that make AI more valuable the longer it runs.

What This Means for Consulting

If AI can generate insights and analysis, does the world still need consultants? The skeptics say no. But AI doesn’t make consulting irrelevant. It raises the bar and changes the work.

Clients don’t just need strategic insights. They need partners who can help them:

• Rethink processes wholesale with AI as the foundation
• Integrate AI into the fabric of operations
• Build the skills and culture that sustain adoption

At West Monroe, we’ve seen this firsthand. When clients try to “buy copilot” or a chatbot of their choice and call it transformation, they fall short. When they commit to reimagining how finance, supply chain, and customer operations actually work, AI produces durable results.

Consulting in this context isn’t about adding more PowerPoint slides to the conversation. It’s about guiding clients—and working with them, side by side to design and build solutions that work across the messy, complex process of adoption and transformation. This is the kind of partnership that actually produces a new kind of ROI.