- AI-enabled finance organisation: finance pioneers already run at 90% automation in reporting and over 80% touchless invoice processing. Isolated pilots do not get there.
- CFO as an orchestrator: the role moves from running the reporting cycle to defining the rules of execution: what agents run, what needs human review, what never leaves finance.
- Live Finance solutions: current, source-traceable numbers and continuous execution replace periodic cycles. Capacity moves from producing the view to acting on it.
Finance is under increasing pressure to deliver faster decisions, stronger control and greater productivity despite growing complexity. The traditional operating model, built around periodic reporting and manual execution, is reaching its limits. Agentic AI changes how the finance function is organised. For CFOs, the question is no longer where to automate, but how to redesign finance around autonomous execution.
Agentic AI in a finance organisation
The word "agent" is doing a lot of work in finance right now. Most of what carries the label is an add-on: an AI feature layered onto existing software that still needs a person to do the actual work. A real AI agent is different. It plans and executes multistep workflows within set guardrails, interprets context, and escalates the exceptions that need a human. It runs the accounting routine continuously, governed and auditable through embedded controls, approvals and source-level traceability. That is the line between point automation and autonomous execution.
The demand is clearly there. BCG reports that 88% of CFOs now call AI essential to their finance agenda, and 96% expect a significant or transformative impact within the next five years.
Yet most finance functions stay stuck at point automation, and the reason is architectural: they add isolated AI features onto legacy systems instead of replacing them. That is what opens the gap to the companies moving to AI-native platforms that run entire finance processes end-to-end through AI agents.
Prioritising AI
~88%
of CFOs rate AI as essential or important to their finance agenda.
Expecting impact
~96%
of CFOs expect significant or transformative impact from AI within five years.
Realising ROI
>25%
ROI already realised by finance leaders that have deployed AI at scale.
Source: BCG, CFO and finance executive survey, 2026.
How is the role of the CFO changing?
Agentic AI calls also for a redefinition of the CFO's skill set. While traditional financial expertise remains fundamental, new skills, particularly in the areas of technology, data and leadership, are gaining strategic importance. The CFO of the future must not only be an excellent financial expert but also an AI-enabled orchestrator of financial execution. Accordingly, CFOs will lead hybrid finance organisations, overseeing human talent alongside the governance of digital agents, AI operating costs, and data integrity.
Competencies of the CFO in the agentic era:
- AI fluency: understanding what agents can and cannot do, so the CFO can set the boundary between autonomous execution and human review with confidence rather than caution.
- Strategic judgement: turning always-current numbers into decisions on capital, liquidity and risk faster than the old cycle allowed. This is the part of the job that grows, not shrinks.
- Process Understanding: understanding of end-to-end finance processes to drive their optimization and redesign in the context of AI.
- Leading a hybrid organization: roles, incentives and retraining, with a clear line on where the function is heading. The CFO must convince and train employees and support their transition to the agentic era.
- Data Literacy: an understanding of data architectures, data quality and data integration, including how data is collected, processed and used by AI agents.
From periodic finance to Live Finance
These new skills describe what finance teams adapt, but they do not fully describe what the CFO is answerable for today as agentic AI also facilitates the transition from periodic cycles to Live Finance. For the CFO, periodic finance meant managing a backward-looking process. Financial work accumulated throughout the month and was consolidated at period end. The CFO had to coordinate submissions, resolve intercompany queries, chase approvals and ensure the close was completed on time. The main bottleneck was therefore often not the accounting itself, but the organisational dependencies surrounding it.
With the rise of automated workflows, financial processes began to run more frequently. Forecasts were updated and controls became more continuous. For the CFO. This meant shorter cycles and fewer bottlenecks. However, the underlying operating model remained unchanged: the period still defined the work.
The current shift is different. Under AI-native Live Finance solutions, the period no longer defines the work: AI agents execute tasks, analyse data, coordinate workflows and handle exceptions continuously across accounting, reporting and treasury. Figures remain current and traceable to source, while decision rights and judgement-intensive exceptions remain with people. The CFO can provide the CEO and the Board of Directors with real-time financial insights and predictive scenarios, improving the quality and speed of strategic decisions. In turn, manual data consolidation, standard reporting, and routine variance analysis become automated by always-on Intelligence.
The operating model shift
| Periodic Finance | Live Finance | |
|---|---|---|
| Team setup | Layered hierarchies across FP&A, accounting and shared services | Lean teams orchestrating workflows and resolving exceptions |
| Scaling law | Output scales with headcount. More volume means more people and more layers | Output scales through automation, with no additional headcount |
| Capacity allocation | Majority of time on data gathering, reconciliation and report assembly | Time available for analysis and business decisions |
| Reporting | Assembled per cycle | Continuously available |
| Variance analysis | Days of manual data stitching across sources | Root cause and ranked corrective options on demand |
| Transactions | Manual entry, matching and approval | Touchless for amounts below the materiality threshold |
| Vendor onboarding | Weeks of sequential checks and manual validation | Days, with exceptions routed to a person |
| Control model | Periodic testing of samples after the fact | Continuous detection of anomalies and prevention in the flow of work |
| Forecast accuracy | Point-in-time, quickly stale | Refreshed against actuals continuously |
The role of the CFO in the era of agentic AI therefore moves from managing the reporting cycle to orchestrating the rules, AI agents, and human decisions through which finance is executed across the enterprise.
