Insurance CFO highlights $80m AI cost saving amid ROI push.
AI cuts costs for customer service.
An insurance sector finance chief has offered a rare example of direct return on investment from AI. In a recent KPMG discussion, the unnamed CFO said the company’s AI-powered virtual agents had reduced inbound call volumes by 20%, yielding over $80 million in annual run-rate savings in its call centres ([1]). Such concrete cost reduction, attributed to automation of routine customer queries, provides a welcome proof point of AI’s tangible value in a specific operational area.
This kind of success story remains the exception. In practice, broad financial gains from AI are elusive for most firms. A mid-2026 global survey by McKinsey found only 37% of organisations could attribute any positive impact on earnings to their AI initiatives ([2]), essentially unchanged from the previous year. The same study showed just 6% of companies qualified as “AI high performers” - defined as those attributing more than 5% of EBIT to AI projects ([3]). The contrast underscores how uncommon it still is to see AI move the needle on the bottom line, outside of well-targeted use cases where metrics like call volume or processing time are owned and tracked by a single team.
Finance chiefs insist on ROI proof.
With economic headwinds and rising AI bills, CFOs are tightening oversight of AI spending. KPMG’s Q2 2026 Global AI Pulse survey - which polled over 2,100 senior leaders across 20 countries - found that many organisations are refocusing from experimentation towards “accountability, AI economics and value”, given that “established ROI remains limited” ([1]). In the same study, nearly half of enterprises reported having rephased or slowed their AI deployments when costs began to outweigh the expected benefits ([2]). In other words, finance leaders are increasingly hitting pause on AI projects that fail to show a clear business case.
CFOs are also demanding better measurement of AI outcomes. One new industry survey of senior finance executives found that only about 22% can fully connect their AI spending to business results at present - highlighting a major gap in ROI governance. This lack of visibility is motivating boards and CEOs to press CFOs for more rigorous tracking of AI value ([3]). Finance chiefs, in turn, are calling for each AI initiative to have defined success metrics and shorter payback periods before committing further investment.
CFO leadership boosts AI payoffs.
The clearest returns on AI tend to emerge when C-level leaders are directly involved in driving the agenda. Research released by IBM’s Institute for Business Value on 30 September found that companies led by “AI-first CFOs” - finance chiefs who take charge of their enterprise’s AI strategy - achieved revenue growth rates 23% higher than peer organisations from 2022 to 2024 ([1]). These AI-focused CFOs are also significantly more likely to report effective execution of company strategy, and they approve funding for new AI initiatives 15% faster than their counterparts ([2]).
The IBM study suggests that strong financial governance and alignment with business goals are critical to capturing value from AI. When CFOs treat AI projects as strategic investments with clear targets, they can drive efficiency gains or revenue growth that shows up in the P&L. In contrast, AI efforts that lack executive ownership or well-defined metrics risk remaining “science projects” that may never scale or deliver meaningful returns.
Trillions in investment, elusive returns.
As investors pour unprecedented sums into AI, the pressure to realise commensurate returns is mounting. Venture capital firm Sequoia Capital estimates that after several years of heavy spending, total AI infrastructure investment in 2026 has reached roughly $1.5 trillion - and that the industry would need to generate about $3 trillion in revenue to justify those costs ([1]). This staggering ROI target underscores the current gap between AI excitement and economic reality. For now, the returns from AI remain concentrated in narrow, use-case-specific successes, while broad-based financial impact across the enterprise is still more promise than practice.