When the Finance Function Comes for Your AI Budget, What Will You Show Them?

Only 20% of organisations are already generating revenue through AI. For the other 80% still calling it an aspiration, the CFO's arrival is not a future risk — it is a present reckonin

Audio summary:

Sign up for my Substack daily AI newsletter here.

See my AI Training course portfolio for corporate Business Leaders here.

Follow me on LinkedIn: https://www.linkedin.com/in/johanosteyn/

For the past three years, AI has enjoyed a kind of protected status inside organisations. Budgets were allocated with unusual generosity. Governance questions were deferred. Pilots proliferated without rigorous business cases. The language of transformation gave leaders cover to invest without having to prove returns. That era is ending. As Fortune reported at the close of 2025, 61% of CEOs say they are under increasing pressure to show returns on their AI investments compared with a year ago. The person exerting that pressure most consistently, most rigorously, and with the least tolerance for aspiration dressed up as strategy is the CFO. And for the majority of organisations whose AI programmes exist somewhere between an ambitious roadmap and a genuine revenue impact, the arrival of that conversation is not a future risk. It is the defining management challenge of 2026 — arriving now, in real time, with real consequences for every AI investment that cannot account for itself.

CONTEXT AND BACKGROUND

The scale of the gap between AI investment and AI returns is striking. Global spending on AI is forecast to reach 2.52 trillion dollars in 2026, a 44% increase year on year, according to Gartner. And yet, as CFO Dive’s analysis of the challenge facing finance leaders makes clear, the question confronting most organisations is no longer whether AI is the right thing to invest in — it is how to actually unlock value and how to measure it. The gap between those two states of the conversation is where most AI programmes currently live. Investment is growing. Accountability is catching up. The distance between them is where the reckoning is happening.

The World Economic Forum has been direct about what the CFO brings to this conversation. The scale of the failure rate inside most AI programmes is more confronting than most boards have been told. MIT's NANDA initiative published The GenAI Divide: State of AI in Business 2025 — a study based on 150 interviews with business leaders, surveys of 350 employees, and analysis of 300 public AI deployments — and its findings were stark. As Fortune reported on the research, about 95 per cent of AI pilot programmes deliver little to no measurable impact on profit and loss.

For South African organisations, the context is sharper still. Constrained capital, high operating costs, and intense margin pressure mean there is no comfortable buffer for AI investments that underperform. The global ROI problem is real everywhere. In South Africa, it is more expensive, less forgiving, and more consequential for the people whose jobs and livelihoods depend on organisational financial health.

INSIGHT AND ANALYSIS

The CFO’s arrival in the AI conversation is changing its nature fundamentally. Where the conversation was previously dominated by technology leaders speaking in the language of capability, possibility, and transformation, the finance function brings a different vocabulary entirely: cost of ownership, payback period, measurable outcomes, and accountability for investment decisions. As one CFO quoted by Fortune put it plainly, the era of buying AI for AI’s sake is over. CFOs will remain willing to invest — but will require clarity on how investment is tied to business outcomes like improved efficiency, productivity, or sustainable growth.

That clarity is precisely what most AI programmes cannot currently provide. Deloitte’s research on the CFO’s role in AI governance identifies three specific challenges finance leaders now face in capturing AI value. The benefits of AI are often intangible — improved customer relations, faster decision-making, reduced risk — in ways that do not map neatly onto traditional ROI frameworks. The technology evolves faster than the metrics created to measure it, meaning that benchmarks become obsolete before they can generate meaningful longitudinal data. And the costs are frequently underestimated, with organisations that budget for model licensing and cloud infrastructure often failing to account for the substantial hidden costs of data labelling, governance setup, and change management.

That last point is where the C-suite misalignment is most acute. Fortune’s reporting reveals that 65 % of CEOs say they are not aligned with their CFO on long-term AI value. The CFO looks at a balance sheet. The business leader looks at a business model transformation. The technology leader looks at innovation capacity and talent. All three are looking at the same AI programme through completely different lenses — and in most organisations, no one has done the work of reconciling those perspectives into a single, coherent account of what the investment is actually returning. When the finance function eventually demands that account, the gap between the lenses becomes the gap between a defended investment and a cancelled one.

IMPLICATIONS

For boards and executive teams, the most urgent action is to establish — before the CFO demands it — a clear, agreed methodology for measuring AI value that goes beyond productivity metrics. The CFO Dive analysis is specific about what this requires: moving beyond efficiency numbers to ask how AI is helping with top-line growth and how it is helping avoid risk and fines. Those are the questions that turn AI from a cost centre into a strategic investment — and they require a level of intentionality about AI programme design that most organisations have not yet applied.

For AI leaders and technology executives, the CFO’s scrutiny is not an obstacle to be managed. It is the governance mechanism that most AI programmes have been missing. The organisations that are generating genuine returns from AI — the 20 per cent already growing revenue through it — share a characteristic identified consistently across the research: they anchored AI initiatives to measurable business outcomes from the beginning, rather than hoping outcomes would become measurable once the technology was in place. That sequence matters enormously. An AI investment designed around a clear business problem with agreed success metrics is a fundamentally different proposition than a pilot that grew into an infrastructure expense without ever being asked to justify itself.

For South African CFOs specifically, the opportunity is significant. The finance function is uniquely positioned to impose the discipline that AI programmes need — not by restricting investment but by insisting that investment decisions are made with the same rigour applied to every other material capital allocation. That means demanding clear business cases before approval, establishing outcome metrics at the point of investment rather than after deployment, and creating accountability structures that connect AI spend to business performance in ways that the board can scrutinise and the organisation can learn from.

CLOSING TAKEAWAY

The question in the title of this article is not rhetorical. It is the actual question that finance leaders are asking in boardrooms across the world right now — and that South African CFOs are beginning to ask with increasing urgency. What will you show them? If the honest answer is a list of pilots, a set of aspirations, and a growing infrastructure bill without a clear line to business outcomes, then the work of building that answer needs to start immediately — not when the question is formally posed. The organisations that will navigate 2026’s AI accountability moment most successfully are not those with the most sophisticated technology. They are those who had the discipline, from the beginning, to treat AI investment the way every consequential investment deserves to be treated: with clarity about what it is supposed to return, rigour about how that return will be measured, and honesty about whether it is actually delivering. The CFO is coming. The question is not whether you are ready for the conversation. It is whether the conversation you have been avoiding is the one that will save your AI programme or end it.

Author Bio: Johan Steyn is a prominent AI thought leader, speaker, and author with a deep understanding of artificial intelligence’s impact on business and society. He is passionate about ethical AI development and its role in shaping a better future. Find out more about Johan’s work at https://www.aiforbusiness.net