How CFOs can secure solid ROI from business AI investments

A recent MIT study that’s roiling the artificial intelligence (AI) world has found that 95% of enterprise AI initiatives fail. This sobering statistic is understandably causing concern among chief financial officers (CFOs) and other leaders.

The report underscores what many finance leaders are already grappling with: While AI promises transformative efficiency, the path to realizing that promise is riddled with complexity. It’s important to note that the study also highlights what the successful 5% are doing right – they’re focusing on human enablement, strategic alignment and disciplined execution.

Whether you’re bullish or bearish on AI, there is no denying it has become a boardroom priority. In a recent McKinsey survey, 78% of organizations report using the technology in at least one business function, up from 55% a year earlier. Generative AI, in particular, has seen remarkable growth. McKinsey found that 71% of businesses are using it regularly, a percentage that nearly doubled in just 10 months.

This underscores the transition from experimentation to execution. AI is now a top-three priority for 75% of C-suite leaders, and finance teams are expected to lead the charge in evaluating its business value.

People are and will remain key components in this value proposition. Cognizant research has found that 90% of jobs will be affected by AI – 52% of them greatly. And a study by the World Economic Forum and PwC shows long-term value emerges when technology is paired with human adaptability and trust. Job augmentation, not just automation, drives sustainable AI productivity gains.

While it may still be difficult to nail down hard-dollar ROI from AI, information on productivity improvements is coming to light.

AI investments typically span multiple cost categories, including technology acquisition, integration with existing systems, employee training and process redesign. And beyond these visible costs, hidden expenses often emerge as a result of data preparation, governance frameworks, integration complexity and adoption support.

Each of these costs can rival the technology’s sticker price. Internally, we evaluate them not just in terms of budget impact, but in terms of organizational readiness and long-term scalability. The productivity gains from AI can be substantial, however, as it reshapes workflows, enables faster decision-making and reduces manual effort.

For example, Cognizant is working with Telstra, the Australian telecoms company, to reimagine the future of engineering using AI agents to help automate processes and workflows. Telstra says it’s already starting to see the potential for gains in quality, velocity and efficiency as a result.

Multi-agent AI systems can be used to power IT operations and enable key business functions like finance and HR to talk to one another through an interconnected system of agents. It’s early days for such systems but initial tests show processes that used to take weeks and involve countless human and technology touch points can be reduced to just a few days.

The productivity gains from AI investment must be reinvested into higher-value work to compound long-term value. From a CFO’s perspective, the real ROI emerges when technology investments are matched by human elements, including skills, trust and time to adapt.

What, then, should financial leaders be asking now about AI’s potential impact? How can they give their organization the best possible odds of success? Asking the following questions would be an excellent start:

AI initiatives often involve upfront costs for model licensing, cloud infrastructure and integration. But other costs, such as data labelling, governance setup and change management, can be substantial. For example, preparing high-quality training data and ensuring compliance with ethical AI standards may require dedicated teams and extended timelines. A comprehensive cost model should include these elements to avoid overruns and set realistic expectations.

Unlike traditional IT projects, AI outcomes can be probabilistic and evolve over time. Establishing cross-functional accountability between finance, operations and data science is essential to monitor performance metrics. A centralized dashboard can help track these KPIs and inform reinvestment decisions, ensuring that AI initiatives remain aligned with business goals.

Don’t chase novelty over substance. AI should be deployed where it can augment human decision-making or automate repetitive tasks with measurable impact. For instance, using AI to streamline claims processing in insurance or accelerate drug discovery in pharma aligns with core business value. Prioritizing such use cases ensures that AI investments deliver tangible returns.

AI can free up employee time, but without a plan to redirect that capacity into innovation or higher-value work, gains may stall. CFOs should work with HR and business leaders to identify areas in which augmented roles can drive growth, such as customer experience, product development or strategic planning. Reinvestment strategies should be explicit and measurable to ensure long-term value creation.

AI adoption often faces skepticism due to fears of job displacement or lack of understanding. Investing in training programmes, transparent communication and inclusive design can build trust and improve uptake. Building AI literacy across the organization ensures teams are empowered to use these tools effectively and responsibly.

Ultimately, AI should be a force for good jobs and human augmentation—a goal aligned with both the World Economic Forum’s Future of Jobs Initiative and Cognizant’s Synapse programme, which will reskill 1 million workers for the AI age by 2026.

Disciplined investment, thoughtful reinvestment and a focus on human enablement will ensure that AI delivers sustainable value across enterprises.