I recently had a conversation with a CFO that I can’t stop thinking about. I asked him how far his organisation had got with AI and automation in finance. His answer was brief.
AI projects, he said, were not even planned in his organisation. Not in preparation. Not under evaluation. Not on the roadmap. Not even planned.
This CFO is not an isolated case. He is a pattern. In the finance organisations I work in, the most advanced AI application is almost always Copilot in Microsoft Office. Sometimes not even that.
Five reasons why AI and automation don’t land in finance
Reason 1: Automation systematically loses out to day-to-day business
Month-end close. Quarter-end close. Auditors. Regulation. The calendar is full before a single automation initiative even begins. The teams that would benefit most have the least time to deal with it.
Reason 2: The processes are not ready for automation
Automation works with stable, documented processes. In many finance departments, processes have grown organically over time, are undocumented and depend on individuals. Whoever automates an unstable process automates the chaos.
Reason 3: The decision chain is too long
Finance. IT. Data protection. Compliance. Executive management. Every body has its own priorities and its own concerns. Without someone with authority and staying power, every initiative dies in coordination loops.
Reason 4: Finance and IT speak different languages
Finance thinks in business processes and audit reliability. IT thinks in systems and development sprints. No amount of goodwill closes this gap. It needs a bridge.
Reason 5: The perfectionism of the complete solution
If a process cannot be automated 100 per cent, nothing happens at all. Yet 70 per cent automation would already be dramatically better than the status quo.
What Copilot really is and what it is not
Copilot in Microsoft Office is a productivity tool for knowledge work. It does not post invoices. It does not reduce manual interventions in transactional finance processes. It speeds up the writing of emails and the summarising of documents.
Confusing these two categories is one of the most common misconceptions I observe in the market. Copilot is not an automation project for finance. It is an office productivity tool.
What a CFO can concretely do now
Step 1:
Identify manual time-wasters. Two hours. An honest conversation with the team. Which tasks recur every week without requiring complex decisions?
Step 2:
Use the SAP potential that has already been paid for. Automated payment runs. Automated dunning processes. Workflow-supported approvals. Many organisations use a fraction of what SAP can already do.
Step 3:
Power Automate for simple automations. Integrated into Microsoft 365. No programming knowledge required.
Step 4:
Data quality as preparation for AI. Every AI application needs good data as its foundation.
Step 5: A pilot project with a narrow scope. A single process. 90 days. Not five pilots at once.
Whether AI lands in finance does not depend on the technology. It depends on the CFO.
A CFO who treats AI as an IT topic gets IT solutions that finance does not use. A CFO who treats AI as a budget line without operational leadership gets expensive projects that fall short of their promises. A CFO who treats automation as an operational leadership task gets results.
What I bring
I am Nicole Vekonj, Interim Finance & Controlling Manager and a Business Automation Manager in training at Haufe. I combine hands-on assignment experience with structured knowledge about process automation that really works in finance organisations.
👉 Submit a project enquiry: zahlenkompetenz.de/en/contact
