Under increasing pressure, how can CIOs convince CFOs to invest in AI?

In 2022, Air Canada tried to put distance between it and its chatbot after a passenger was told he could get reimbursed for a bereavement fare. After a legal battle, Air Canada was ultimately ordered to honor the chatbot’s promise but the chatbot was taken offline as a result.

Examples of AI-powered tools going rogue or being ineffective have been quite common in the past few years as companies jumped on the hype. But just like Air Canada’s chatbot, so far, the wave can sometimes promise more than it can deliver, which worries CIOs and CFOs.

When it comes to AI investments, companies want results — and they want them fast. Half of CFOs say they plan to cut AI funding if it doesn’t show measurable ROI within a year, according to a global survey from accounts payable automation firm Basware, which included 400 CFOs and finance leaders.

While 80% of organizations plan on increasing AI investment, CFOs often don’t know how to start. About a third of them feel they lack a clear vision.

For CIOs, this means they must go above and beyond to bridge the gap between AI’s potential and the CFOs’ demand for quick, tangible, and reliable results. However, this is often easier said than done. Being strategic about AI and measuring whether those investments are paying off requires clear goals, reliable data, and collaboration — challenges many organizations struggle to overcome.

Around 60% of global CIOs believe that increased revenue alone justifies the cost of AI, and a similar proportion says time savings are sufficient to validate the investment. However, only around a third of them actively measure both, according to a recent study by revenue intelligence leader Gong, which surveyed over 500 IT leaders and CIOs across the US and UK.

CIOs are under pressure to validate AI investments and assure CFOs of a clear path of implementation that will ensure ROI. This requires not only selecting the right projects but also clearly defining how success can be measured.

How to know what to prioritize

AI has made remarkable strides over the past year, but its adoption has also uncovered a host of shortcomings like dangerous hallucinations and expensive implementation.

Lucidworks’ study of gen AI investment says that in 2024, business leaders are slowing down spending to balance the benefits, costs, and risks of this relatively new technology. Also in 2024, 42% of companies reported that their gen AI initiatives have yet to deliver meaningful results.

While businesses see the potential of AI, they often remain cautious when it comes to investing, weighing the risks and costs, so picking the right projects can be challenging. To maximize the impact of AI initiatives, organizations should focus on aligning each project with their overall corporate strategy and long-term goals.

“Start by identifying where AI can enhance core capabilities, whether it’s improving product quality, accelerating time-to-market, or enabling data-driven decision making,” says Karli Kalpala, head of UK and Ireland, and strategic transformation at Digital Workforce.

Prioritizing scalability is also critical, Kalpala adds. “Build for scale early on and ensure each department can adopt and adapt the tools with minimal friction,” he says. “Everyone can build one AI solution, but those who will transform their organization with AI need to build and maintain hundreds of AI solutions across the whole organization.”

Kristen Costagliola, CTO at Syncro, suggests applying Harvard professor Michael Porter’s strategic competitive levers, which identify three primary approaches a company can use to achieve competitive advantage: cost leadership, differentiation, and focus.

“This perspective helps CIOs think more strategically about where to apply AI within their businesses,” she says. “For example, if a company’s strategy is cost leadership, the CIO would prioritize projects that drive efficiency to lower costs. Alternatively, if a business is looking to provide a differentiated product or service, a CIO would look to apply AI to innovate and stand out from their competitors.”

Organizations already generate large volumes of high-quality data in some areas and have well-defined pain points. Perhaps they should start there, says Om Moolchandani, the co-founder, CISO, and CPO at Tuskira.ai.

“Customer experience optimization, supply chain forecasting, demand prediction, and preventive maintenance tend to yield quick wins,” he says. “By focusing on domains where data quality is sufficient and success metrics are clear — such as increased conversion rates, reduced downtime, or improved operational efficiency — companies can more easily quantify the value AI brings.”

It’s also important to understand the priorities CFOs have. According to Basware’s survey, 75% of CFOs prefer AI investments to focus on areas like e-invoicing compliance and regulatory requirements.

But choosing the projects to prioritize is no simple task and underscores a crucial point: businesses are eager to see fast, tangible impact from their AI initiatives. However, validating AI investments and quantifying that impact is often two different matters.

Substantiate a project, measure impact

Every AI initiative needs to be validated before full-scale implementation, and this often requires a combination of technical and strategic assessments. Companies need to focus on goals, testing, and people in their effort to determine if an AI project is viable.

“The first step is to define the metrics,” says Scott Willson, a tech evangelist at ServiceNow’s multi-instance management platform xtype. “Are the KPIs aligned with measurable business outcomes that stakeholders can rally behind?” Clear metrics not only guide the project but also help communicate its value to decision-makers across the organization.

The next step is prototyping and piloting the project using a minimum viable product. This helps test assumptions, gather valuable insights, and refine the solution before full deployment.

Kalpala also suggests testing the AI solution through a pilot program in real-world business environments. “These tests help evaluate whether the AI solutions integrate well with existing systems and workflows, or if major system overhauls are required,” he says. It’s also important to identify challenges like data privacy, technical readiness, and organizational change management.

To foster productive discussions with CFOs and other leaders, CIOs can present a clear set of early success metrics, such as workflow automation or error rate reduction. Using business language and explaining how these metrics directly align with organizational goals can help stakeholders understand the benefits of AI projects, and can make them more likely to support them.

It’s also critical to ensure strong stakeholder buy-in by engaging cross-functional teams, including IT and business leaders. “When everyone is aligned, you minimize risks and potential delays, and set the stage for success with the project,” Willson says.

With half of CFOs in the Basware survey saying they plan to cut AI funding if it doesn’t show measurable ROI within a year, the pressure on CIOs to deliver has never been greater.

“The key to ensuring AI projects deliver ROI quickly is starting small,” Willson adds. “Break the project into manageable, experimental phases to learn and adapt quickly. Focus on use cases that provide quick wins and tangible results, so the company sees immediate benefits.”

Ala Shaabana, co-founder of decentralized ML network Bittensor, suggests companies focus on tackling low-hanging fruit first, those projects that are straightforward to implement and deliver results quickly.

“Start with targeted use cases that address high-impact, low-complexity problems such as process automation or customer support optimization,” he says. “Leverage existing data and infrastructure to avoid costly delays in data collection or system integration. And ensure continuous monitoring and iteration to quickly address issues, sustain momentum, and maximize returns.”

But one problem CIOs face is the lack of good benchmarks for AI ROI. “This presents a significant challenge because AI’s value is often multifaceted and evolves over time,” says Eric Helmer, CTO at Rimini Street. “Organizations should also establish their own baseline metrics for pre-AI performance to track improvements.”

Kalpala agrees and adds that every AI deployment is unique, and success metrics need to be tailored to fit the specific industry, use case, and organizational needs. “For insurance or financial services enterprises, benchmarks should focus on reducing manual errors, improving service speed, and reducing direct expenses to deliver the core services,” he says.

Strategies to convince CFOs

Securing the support of CFOs can be a tough challenge when implementing AI initiatives. However, once CIOs are confident they’ve chosen the right project, there are several steps they can take to strengthen their case and gain approval.

“When it comes to convincing CFOs of AI’s potential, it’s crucial to present a strong business case,” says Willson. “Tie the AI project directly to strategic goals like boosting revenue or reducing operational costs.”

Moolchandani recommends translating technical outcomes into financial numbers. “Present expected gains or savings in tangible terms, such as a projected percentage increase in upsell opportunities, or a dollar amount saved through automated processes,” he says. He also recommends showcasing rapid prototyping and real results, which can instill confidence.

The implementation plan should be documented well, adds Syncro’s Costagliola, and should include clear milestones. “Clearly outlining the time horizon and the expected time for results is paramount to convincing the CFO and gaining buy-in,” she says. The impact on the organization should also be presented in detail.

Willson also emphasizes the idea of de-risking the project, which includes implementing quality control checks along the way. “This will ensure alignment with business priorities and compliance standards, and make the investment more attractive,” Willson says.

Kalpala points out that CIOs need to think beyond the next quarter or year and explain how AI implementation can transform the business. “Explain the strategic, long-term impact of the project by demonstrating its ability to drive innovation and workforce transformation,” he says. “We’re moving from software as a service, to service as a software. Use this opportunity to show what this shift means for the organization in terms of new opportunities and pricing strategies.”