CFOs Use AI for Efficiency, Not Growth—Trust Gap Persists

While many companies are progressing well in using artificial intelligence tools for productivity gains, far fewer are employing them to help with strategic growth initiatives. There is a lack of trust — yet overcoming that barrier may separate leaders from laggards in the upcoming years, EY-Parthenon contends in a new report.

The data comes from a survey commissioned by EY-P, the consulting arm of Ernst & Young. Of the 271 participants — all at the vice president level or higher and charged with overseeing company growth — 63% said AI is helping them with efficiency and productivity. At the same time, only 14% of those polled said they are leveraging AI to stay ahead of competitors, while just 8% are doing so to reach new customers and 7% to diversify revenue streams.

That's a 49-point gap between what companies are doing with AI and what they could be doing. The numbers don't lie: executives are comfortable with AI as a cost-cutting tool but hesitate to let it drive expansion.

EY-P suggested in a research report that the need for change should be obvious, considering that 80% of the corporate growth leaders said the environment for business growth is more challenging than it was a year ago. In fact, almost all (97%) of the survey participants said external forces have led their business to change its growth strategy in the past year.

"The survey reflects what is in the news every day," EY-P wrote. That is, the leading drivers of changing growth strategies, the report said, are: (1) geopolitical and economic pressures and volatility, affecting 73% of survey participants, and (2) technological innovation, influencing 58% of those polled.

"The days of a three-year growth plan may be a thing of the past, and companies are looking for a catalyst to accelerate expansion despite the current conditions," the report declared. "Enter AI for growth, not just productivity."

Many survey participants are optimistic about possible improvements in selling or serving customers (63%) or developing new growth markets (57%). However, a troubling perception gap remains. While a large majority (78%) of those surveyed said they expect AI to accelerate their company's growth rate, just 34% said they trust AI to support decision-making in growth-related areas.

That 44-point spread between expectation and trust is where the real story lives. Companies want AI to work for them, but they won't let it drive the car (they're still gripping the wheel, even when the autopilot is humming along).

The leading factors preventing companies from innovating faster than competitors are risk and compliance challenges and legacy technology and infrastructure. These aren't abstract concerns. They're the physical reality of trying to integrate new AI systems with decades-old ERP platforms, legacy databases, and compliance frameworks that weren't designed for machine learning.

According to Mitch Berlin, EY Americas vice chair for EY-P, the companies that will "win" will be "those that figure out first how to leverage AI to drive growth by innovating faster, hyper-personalizing their offerings, and launching new products and services."

EY-P said working with clients on the frontier of leveraging AI for growth has revealed a few key emerging themes. First, harnessing AI to mine a company's intellectual property and patent data to help drive growth. "Companies can then find new uses for products that can be developed and markets that can be addressed."

Second, driving greater value from core systems. "Recent advances in AI-based reasoning allow companies to separate decision intelligence from transactional systems … and continuously [learn] from outcomes rather than relying on static rules."

Third, tapping into the promise of neuro-symbolic AI. "NSAI combines the ability of neural networks to learn from structured and unstructured data with a symbolic layer of rules and parameters that create consistent, auditable and transparent results for decision-making."

The neuro-symbolic approach is particularly interesting for finance and compliance-heavy industries. It addresses the core trust issue by combining machine learning's pattern recognition with rule-based systems that auditors can actually understand. That's not just a technical detail — it's the difference between an AI system that gets deployed and one that sits in a sandbox forever.

Independent reporting from CFO.com corroborates the timeline and scope of the changes. The publication's coverage of the EY-P survey provides the primary source for these statistics and executive commentary.

Additional context from a 2026 Kyriba survey shows the trust gap extends beyond growth decisions. In that separate study, 77% of CFOs cite security and privacy as critical risks. Only 47% have integrated AI into some processes, and just 45% use it in the majority of decision-making. The pattern is consistent: optimism is high, operational readiness is not.

The physical reality of AI deployment matters here. When a finance team implements an AI agent, they're not just clicking a button. They're mapping workflows, defining thresholds, setting up audit trails, and establishing review cadences. Every action the agent takes — what it did, why, what data it used, what policy it applied, and what it escalated — has to be logged in a form that's accessible to the team and understood by auditors.

That's where many pilots fail. The "human-in-the-loop" label often means humans are still doing the actual work while AI prepares tasks or organizes queues. Control should mean review, approval, and auditability — not permanent manual execution. That distinction is where AI pilots either start creating capacity or become another operational burden.

The test is whether AI changes who does the work. If your team is still verifying every input, completing every action, clearing the same exceptions every run, and manually documenting every decision, the system has preserved the old operating model. It may make the queue cleaner, but it hasn't created meaningful capacity.

The better approach moves people out of constant execution and into control. AI completes the assigned work inside defined scope, thresholds, permissions, audit requirements, and review cadence. The finance team reviews the results, approves material actions, investigates exceptions, and gives the agent feedback so it handles more cases on its own over time.

Whether users actually pay for AI-driven growth remains the real question. Companies can cut costs with AI tools, but convincing boards to let algorithms drive revenue strategy requires trust that most organizations haven't built yet. The technology exists. The willingness to deploy it at scale does not.

Time will tell if the 34% who trust AI for growth decisions become the 34% who dominate their markets. For now, most executives are content to let AI handle the boring stuff while humans keep steering the ship. That's a safe strategy — until a competitor decides to let the autopilot take over completely.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn