As AI transforms finance, CFOs must stay ahead
AI in finance is past the pilot phase. According to the latest CFO sentiment data, 75% of finance teams now report either extensive or partial use of AI, up from 30% just a year ago. The real question: can your finance function keep pace with the rest of the enterprise?
Across the broader organisation, 93% report using AI-up from 54% at the start of 2024. "Extensive" adoption has grown from 3% in the first half of 2025 to 12% in the second. While few claim full maturity, capability is building fast across data, tech, and operations. The bottleneck isn't tools-it's organisational design, skills, and culture.
The gap: confidence in tasks, hesitation in judgment
Most teams started with automation-data capture, validation, reconciliations. That's fine for quick wins. Now, the conversation has shifted to AI-led analysis, insight, decision-making, and end-to-end processing.
CFOs trust AI for structured work: 71% expect transformational or material value in technical areas like data processing and validation. But confidence drops in higher-judgment areas-only 41% expect similar gains in planning, budgeting, and forecasting. That hesitancy is understandable, but it's also where competitive advantage will be won.
Humans + AI: where finance creates real leverage
AI works best when it amplifies people, not replaces them. Consider capital planning models: predictive engines generate fast, accurate baseline outputs. Add human scenario design, constraints, and strategic nuance-and you get insights that challenge assumptions and move the business.
The same pattern holds for analysis. AI can generate a solid first pass, flag anomalies, and surface drivers. Finance professionals turn that into direction, trade-offs, and decisions. As agentic AI becomes common in enterprise tools, expect less manual number work and more time spent on what actually shifts value.
What top CFOs are doing now
- Run a readiness check across four domains: data, tech, process, and culture. Fix the weakest link first.
- Prioritise 2-3 end-to-end use cases (e.g., close-to-report, working capital, FP&A scenario design) with clear outcome metrics.
- Stand up lightweight governance: model inventory, validation standards, and approval workflows.
- Invest in targeted training for analysts, controllers, and FP&A to reduce tool anxiety and raise judgment quality.
- Pilot agentic workflows in planning, forecasting, and variance analysis with tight human oversight.
- Tighten data foundations: master data, lineage, access controls, and API-based integrations.
- Create a cross-functional AI council (Finance, Data, Risk, Tech) to remove blockers and share wins.
- Measure value weekly: time saved, error rates, forecast accuracy, and decision cycle time.
A 90-day execution plan
- Days 0-30: Audit current use, pick the use cases, define metrics, establish governance, and prep data.
- Days 31-60: Build and test pilots, run change management, and train "super users" in each team.
- Days 61-90: Expand to adjacent processes, publish results, refine controls, and lock in operating norms.
Upskill the team without the noise
If your biggest barrier is capability, start simple: hands-on courses for finance roles and an up-to-date view of AI tools built for accounting, FP&A, and treasury. Two useful places to begin:
- AI courses by job - curated learning paths for finance teams.
- AI tools for finance - vetted tools to speed analysis and controls.
The takeaway
AI won't replace finance, but finance leaders who don't update culture, skills, and processes will be replaced by those who do. Set clear milestones, pair humans with AI on work that matters, and build the muscle to scale. The teams that act now will define the standard for how finance operates over the next decade.
Your membership also unlocks:
