AI won’t save you. But your data might

As companies race to capture ROI from AI, many are learning a hard truth: AI only creates value when it’s anchored in business outcomes and built on strong data foundations. Jeremy Bruck, a partner at global business and tech consulting firm West Monroe, which advises private equity firms and their portfolio companies on how to drive measurable results from AI, helps companies cut through the noise and focus on where data truly drives differentiation.

I caught up with him to discuss how boards, CIOs, and CFOs can target their AI investments for real impact.

Why are we all talking about AI investments right now?

Because companies are finding that AI is somehow simultaneously ubiquitous and elusive. It’s everywhere in conversation, yet difficult to translate into tangible business value. Just as companies realized that getting a mobile app wasn’t a digital strategy, they’re recognizing that “doing AI” is more than licensing frontier models and training their teams on prompt engineering.

Then how should boards advise their management teams?

Boards should stop asking how to use AI and start asking where their data, powered by AI, can create value.

The recent MIT study citing that more than 90% of AI projects stall before scaling got a lot of attention, but the reason for this failure isn’t that the technology isn’t ready. Instead, it’s because organizations fail to focus on the right opportunities, provide the right context engines to power AI solutions, and appreciate the change management required for success.

Management teams need to ask the same questions they would of any major initiative: How will this generate revenue, expand margins, or mitigate risk? If you can’t tie your AI investment to a KPI and estimate its value, it’s not ready for funding.

How can CIOs separate shiny AI objects from real value?

Start with a business case, not a model. Define the KPI, impacted users, expected actions to realize value, and timeline to value. Expected actions are critical, as increased efficiency doesn’t result in realized value unless you increase demand or reduce your cost basis. Also, don’t assume the tools will differentiate you; we’re already seeing convergence across solution providers. Context engineering is also critical as AI solutions require the appropriate context to drive consistent and accurate outputs. The real advantage lies in where and how you apply the tools to drive outcomes.

Why is AI being treated differently than past emerging technologies?

When CRMs, ERPs, and data warehouses emerged, employees had no personal context for them. But everyone has now interacted with ChatGPT or another AI tool. People extrapolate their positive personal experiences with AI and assume it’ll just as easily translate to business scenarios. Executives are expecting an AI easy button. Unfortunately, tinkering with a recipe isn’t the same as transforming a business process. Enterprise AI requires context engineering, iteration, and an expectation of change across an organization.

How should CFOs think about AI investments?

Early AI adoption was driven by a combination of hype and FOMO. Now, CFOs want business results, not experiments, from AI investments. The challenge is that no one knows which vendors or models will win, or even exist, in three years, so investing in AI solutions presents risk. That’s why the smartest investments aren’t about specific tools but in data infrastructure and modular architectures that promote flexibility and data sovereignty. The winning strategy in 2025 isn’t picking an AI vendor, it’s building an operating model that lets you pivot as the market evolves.

So, it’s really data maturity that drives AI ROI.

Exactly. If everyone has access to the same foundational models, the differentiator becomes your proprietary data and how effectively you can use it. Data is the asset and AI is the application. Without strong data quality, governance of and access to your pilots won’t scale. CFOs should view investments in modern data platforms and governance as the foundation for any AI initiative.

One of our clients, for example, recently stood up a tiger team to build gen AI-powered workflows in less than six weeks. Their secret wasn’t in the AI solutions they selected but that their six weeks was really based on months of disciplined investment in data governance and infrastructure, which laid the foundation for them to move at the speed of AI.

What do we mean by data liquidity and why does it matter?

Data liquidity describes how available, high-quality, and extensible your data is. Start by ensuring your first-party data is strong, then supplement with third-party data to enhance insights. Today, it’s easier than ever to enrich your datasets through marketplaces and secure data exchanges, which lets companies combine proprietary and external data responsibly. The goal is to complement what you have, not be limited by it.

How can CIOs and CFOs overcome the common barriers to starting?

Don’t let perfect get in the way of good. You don’t need a perfect ROI model to begin. Do enough work to identify the biggest opportunities, then deploy quickly and iterate to capture value.

Also, don’t overlook governance. Involve legal and compliance early and engage them as partners, not blockers. When companies fail to define AI guardrails upfront, they often limit themselves with incorrect assumptions about what compliance will and won’t support.

Cross-functional collaboration, clear data ownership, context-powered solutions, and a focus on pragmatic value creation are what separate leaders from laggards.