Artificial intelligence is reshaping business operations, so CFOs face a critical challenge: traditional financial metrics are becoming obsolete. Successfully meeting this evolution means a fundamental reimagining of how finance leaders define success, allocate resources, and communicate value.

"We're moving from reporting what happened to predicting what will happen," says Ying Miao, CFO of Converge, a performance marketing firm with annual sales of over $100 million. "Boards require more than just past-reporting. They need future-looking insights that explain 'why' so we can plan better."

Supporting this change is Helix, the company’s proprietary AI-powered business intelligence system, which analyzes data across channels, down to individual customer interactions. Working with Converge’s data engineering professionals, the company’s Finance department built a database with a chat interface that allows stakeholders to ask questions across data dimensions.

“With this development, the CFO's role shifts from gatekeeper to enabler,” Miao notes. “This ensures insights flow to decision-makers while maintaining appropriate controls.”

Key performance indicators are also changing. While EBITDA and profit margins remain important, “they capture only a fraction of AI's business impact,” Miao explains. ”Real value often lies upstream, in lead generation, pricing optimization, and sales efficiency that traditional accounting struggles to quantify. It's not just about dollars but creating organic efficiency. Can we hire less or do more with the same number of people? That's the question boards want answered."

Converge has developed KPIs around such metrics as sales velocity, pricing models, lead generation quality, and workforce productivity. The company tracks whether AI tools enable scaling without proportional headcount increases, critical for growth-stage companies.

This expanded framework requires CFOs to evaluate AI's impact on earlier marketing funnel stages: targeting precision, creative production acceleration, and enhanced customer segmentation.

The challenge is capturing this latent value without focusing on the wrong metrics. "We used to look at gross margin only, but that's fake," Miao notes. “Our team now tracks comprehensive metrics, including onboarding costs and human capital investment, measuring time to earn back sales and labor costs.”

To ensure the right growth path while maintaining security, the company’s governance includes board-level oversight and a dedicated AI committee. Miao says this is particularly important for companies that handle proprietary information and client data.

"It's a bridge between the CEO and CFO," she explains. "AI may not generate immediate profit, it might be a cost center first, but it leads to stickiness. When you provide better insights and tools, clients are more likely to stay with you."

AI’s long-term returns include broader access to financial insights across departments, Miao says. “The challenge has always been democratizing analytics without sacrificing data security or integrity. When I joined Converge about two years ago, financial information was siloed at the top. Information on financials was only available to top management, the board, or investors. Employees who needed data had to email Finance to ask questions."

To replace that cumbersome structure, her team instituted monthly all-hands meetings, weekly standups to discuss financials, and distributed explanations of why certain metrics matter. This enabled Converge to move from a top-down forecasting model to a streamlined one featuring collaborative monthly reforecasts that incorporate input from account executives and business development teams.

But the increased adoption of AI means CFOs have to be ready to adopt a new mindset, since assessing AI ROI requires a different framework, compared to traditional software evaluations.

To help benchmark results, Converge surveys employees, and tests competing approaches. “Every employee has Copilot Premium access, allowing comparison across different AI platforms,” says Miao. “This experimentation extends to external data integration, combining internal Helix data with external market signals to identify causal relationships. AI lets us leverage big data in ways we couldn't before. This capability to correlate disparate data sources represents a qualitative shift from traditional analytics.”

A Fork In The Road -

Once a decision has been reached to adopt AI, a CFO must consider whether to build proprietary AI tools, partner with platforms, or license off-the-shelf solutions, according to Miao.

“There are three areas we examine,” Miao relates. “What goal do we want to achieve? Do we need better insights for clients; or for our own operations; or better, faster creative, and why? That helps us identify whether it's better to buy tools or develop them internally."

The answer is not an across-the-board one. “Certain applications demand internal development, she says. “For example, our Helix intelligence system handles proprietary information and personally identifiable information subject to marketing regulations, so we developed it in-house.”

Conversely, for financial forecasting without proprietary algorithms, Converge uses external tools like Gemini. "I don't want to develop forecasting on our own, since it's not proprietary to our competitive advantage," she notes.

Miao says that the transformation CFOs face isn't optional, it's existential.

“As AI reshapes business operations, finance must evolve from scorekeeper to strategic navigator, from reporter of history to predictor of futures.”

Fundamentally, she adds, it requires CFOs to see themselves differently, “We are no longer guardians of caution, but must be champions of intelligent risk-taking. Not as data gatekeepers but as insight enablers.”

“For those who hesitate, the risk isn't just falling behind on technology adoption, “ Miao warns. “It means losing relevance in the C-suite as AI transforms the very nature of business strategy. You must be able to speak the language of value creation in an age when data, prediction, and continuous adaptation matter more than ever before.”