According to Deloitte’s latest CFO (Chief Financial Officer) survey, nearly nine out of ten CFOs expect human work to remain predominant in their function three years from now. On 23 October 2026, the magazine PME published Alexandre Buga’s analysis of artificial intelligence (AI) for SMEs, small and medium-sized enterprises: the tool should take the load off the CFO, not replace them.
What Deloitte’s CFO survey says
The PME article first highlights one finding of the survey: humans will stay in control of the finance function. Nearly nine out of ten CFOs believe this will hold over a three-year horizon.
Their role will nonetheless change. They will spend less time processing financial information and more time interpreting it and making decisions. AI will handle larger volumes of data, but finance managers will keep control of the decisions.
Swiss CFOs also name the areas where they plan to invest in AI:
- data management and analysis;
- operational efficiency;
- planning and reporting.
What is the expert’s opinion
The rest of the article is Alexandre Buga’s reading, not a survey finding. In his view, investing in AI does not necessarily mean savings. The financial return is not immediate, and the investment must first create value for the company.
The author also draws on another study, the swissVR Monitor, which he presents as “our study”. Many companies in it report having a governance framework for generative AI. But fewer than one in five have all content validated by a person. For him, a directive on the use of AI is therefore not enough.
Which finance tasks can AI automate?
Processing supplier invoices is the first example given. AI extracts the data and spots duplicates, unusual amounts or changes to bank details. It then flags these cases to the person responsible, who checks them.
Cash-flow forecasting is the second example. AI analyses historical flows, open invoices, payment habits and operational data. From these it draws several scenarios, useful in the face of fluctuating orders, late payments or volatile costs.
These examples describe what the tool can do. The article does not say how many Swiss SMEs already use it this way.
What remains human in corporate finance
Final approval of a payment remains subject to the usual controls. The aim, according to the author, is not to eliminate human intervention but to concentrate it where the risk is highest.
Forecasting is not deciding. Management assesses the scenarios and adds what the systems do not know, for example a customer announcing a restructuring. It then chooses to preserve liquidity, adjust spending or seek financing.
The CFO thus keeps what cannot be delegated: judgement, dialogue with stakeholders and the quality of decisions. For Alexandre Buga, the point is to automate what can be automated without handing responsibility to a machine.
Survey or opinion: who says what
| Point | Source | What is said |
|---|---|---|
| Role of humans | Deloitte CFO survey | Nearly nine out of ten CFOs consider it predominant in three years |
| Investment priorities | Deloitte CFO survey | Data, operational efficiency, planning and reporting |
| Human validation of content | swissVR Monitor | Fewer than one company in five has everything validated by a person |
| Profitability of AI | Author’s opinion | No guaranteed savings, value created comes first |
| Invoices and cash flow | Author’s opinion | Examples of use, with human control of anomalies and decisions |
AI for SMEs: the first concrete steps in Switzerland
Many AI projects remain at the experimental stage, according to the author. The difficulty is integrating them into processes, controls and responsibilities. An OpenAI report on AI agents in European SMEs likewise called for moving beyond pilot projects.
To move from trial to use, Alexandre Buga proposes a simple approach:
- appoint a person responsible for the project;
- set the costs, the expected results and the criteria for measuring them;
- decide which tasks are automated, which exceptions are checked and who is accountable for them;
- evaluate the tool after a defined period, then adapt it or stop it.
The criteria that count are time saved, errors avoided and better decisions.
Data is the other workstream. An SME must comply with Swiss data protection legislation, including when choosing the tool. Before passing on data about customers, suppliers or employees, it must know where the data is processed, for what purposes it may be reused and who can access it.
The author therefore recommends suitable professional tools. He also advises simple rules on which data may, or may not, be used in them.
What we do not know yet
The PME article gives neither the address of the CFO survey, nor its date, nor the number of CFOs surveyed. The exact percentage behind “nearly nine out of ten” is not published either.
Also unknown are the share of Swiss CFOs who have already automated a task, and the measured gain in an SME. The available figures describe expectations and investment intentions, not results achieved.
