CFOs back AI to transform finance - but the trust gap is widening
Two things can be true at once: AI will redefine finance, and you still can't trust it blindly. New research from Kyriba shows 67% of CFOs expect AI to drive the biggest shift in their role over the next five years - up 14 points in six months. At the same time, 77% call privacy and security critical risks. Adoption is accelerating, but confidence in control isn't keeping pace.
The signal in the numbers
- 67% expect AI to be the top driver of role transformation over five years (+14pts vs. six months ago).
- 47% have already integrated AI into core processes.
- 77% rate privacy and security as critical risks.
- Operational priorities: AI adoption (53%), data reliability (31%), security and fraud prevention (27%).
- OPR Index (0-200 scale): Global score 93.28 ("measured confidence"). Components - business outlook 77.43, operational readiness 90.21, external pressures 74.36.
- Regional leaders: Singapore, Germany, UK, US (higher preparedness, lower perceived risk). More cautious: France, Italy, Spain, Japan.
As Kyriba's Chief Product Officer, Monica Green Boydston, put it: CFOs are optimistic about AI's potential, but they're implementing it with strategic caution around security and privacy. That tension is the point. Ambition without guardrails is just risk.
Why the trust gap persists
AI raises the stakes on data handling, vendor exposure, and model behavior. The attack surface expands; regulations keep moving; outcomes can be opaque. Finance leaders are expected to deliver efficiency and insights without introducing brand, compliance, or cash risk.
If you need a reference playbook, the NIST AI Risk Management Framework is a solid starting point for governance and control.
What high-performing finance teams do next
- Set clear guardrails: Define approved AI use cases, risk appetite, and decision gates. Separate experiments from production.
- Fix data first: Establish lineage, quality SLAs, and PII rules. Limit sensitive data exposure. Log everything.
- Tighten security: Enforce least privilege and strong identity, encrypt inputs/outputs, apply DLP, and run vendor due diligence (SOC 2/ISO 27001).
- Stand up model risk management: Inventory models, validate regularly, check bias, add human-in-the-loop for material decisions, and red-team critical workflows.
- Start with low-risk, high-ROI pilots: AP anomaly detection, cash forecasting, reconciliations. Track impact by forecast accuracy, DSO/DPO, fraud loss rate, and close cycle time.
- Increase cadence: Rolling forecasts, scenario planning, and tighter variance reviews to speed response under volatility.
- Upskill the team: Pair finance expertise with AI literacy and data fluency. For a curated starting point, see AI tools for finance.
Growth and resilience: what CFOs are actually doing
Finance is shifting away from "growth at all costs" toward operational discipline. Globally, 37% are increasing the frequency of forecasting updates, and 31% are rebalancing debt and capital structures to manage exposure and the cost of capital.
- US: More reporting and forecasting activity.
- Singapore: Emphasis on balance sheet changes.
- UK and Spain: New software and automation adoption.
- Japan and Germany: Restructuring and treasury automation.
Risk management is getting more analytical
Finance teams are moving from "avoid risk" to "model and measure it." 61% use scenario planning, 57% use AI-powered analytics, and 37% are increasing forecasting frequency. Faster cycles, tighter controls, better decisions.
UK snapshot
The UK stands out with a more upbeat outlook: 82% report a positive view of the economic environment. AI is embedded: 72% see it as a key driver, and 95% say it's already being integrated. Another 95% believe they're ready to handle upcoming macro changes, while interest rates, inflation, and market volatility remain top concerns.
Board conversation: keep it simple
- Strategy: Where AI creates measurable value in the P&L and balance sheet.
- Controls: Data, security, and model governance that cap downside risk.
- Metrics: A short scorecard (forecast accuracy, working capital days, fraud losses, cycle time) tied to incentives.
Bottom line: treat AI like any other financial asset. Allocate capital with clear guardrails, measure outcomes relentlessly, and scale what works. The teams that do this will move fast without breaking things that matter.
Your membership also unlocks:
