5 Steps CFOs Can Take to Maximize ROI From AI Initiatives

Successful AI implementation starts with a clearly defined business problem or objective. Don’t simply adopt what’s worked elsewhere or start from the latest technology. Instead, look for clear alignment between a use case’s scope and your organizational priorities, and then evaluate it against expected value and feasibility. The same use case can offer wildly different potential from one organization to the next. By tailoring AI opportunities to your company’s unique context, maturity, resource capabilities and goals, you’ll be more likely to achieve positive, impactful results from the pilots you select. 

Low data literacy and technical skills in finance are a major stumbling block for AI adoption. Unlock AI’s full value by amping up the finance team’s foundational data science and generative AI (GenAI) skills — such as basic Python, data wrangling and modeling, prompt engineering and GenAI code generation. At the same time, start developing citizen data scientists within the function, who blend technical proficiency and strong business acumen, instead of relying on professional data scientists who often lack functional context. Making this shift could address as much as 90% of finance’s AI opportunity.

Improving team skills will allow more finance staff to work alongside AI in shared roles, which boosts team efficiency and frees staff to focus on higher-value tasks. By 2029, Gartner predicts that one-third of finance staff will hold “shared jobs,” in which AI and a person are responsible for performing a single job role together.

The fastest starting point for AI value is to take advantage of embedded functionality in your existing software. Gartner insights reveal that 99% of finance teams intend to “buy” AI capabilities from third-party vendors — but 60% of finance buyers have experienced postpurchase regret from recent software purchases.

To go beyond vendor promises, architect your vendor engagements to ensure vendors fully understand your organizational processes and key functional capabilities. Explore a variety of vendors and use innovation-based metrics to evaluate performance.

Balancing the cost of AI relative to value is inherently challenging, as organizations must continually determine if the increasing amounts going in are worth the payoff. In some organizations, initial AI project cost estimates have been off by as much as 500% to 1,000%, with ongoing costs proving especially volatile and fast-moving. Look closely at key drivers of cost volatility — for example, usage, varied vendor pricing, or data management demands for the AI model — and push for greater transparency in vendor pricing.

Proactive value management is key to delivering sustained ROI. Develop consistent methods for measuring and discussing AI’s value, a crucial step for evaluating use cases across the enterprise and accelerating funding for the most promising pilots. Enterprises that manage AI investments with a portfolio approach are more than twice as likely to reach mature levels of AI implementation.

AI adoption introduces new and evolving risks related to data privacy, security, intellectual property and AI model performance (such as hallucinations and bias) that must be balanced against potential value in a sound AI strategy.

CFOs should champion an AI governance framework for finance and the enterprise that centers around data integrity, continuous model testing, and human oversight. Lean on your expertise and experience in financial oversight and risk management to take a phased approach that aligns with your organization’s current AI maturity while supporting an accelerated value strategy.