New data from the SAP Concur annual CFO Insights Survey has revealed how finance leaders are implementing AI. Findings suggest finance is cautiously integrating AI into forecasting workflows, helping analyse risk, cash flow, and expenses. These are the top insights:
Cost forecasting is a top-billed issue
Inaccurate forecasting is ranked as a top three internal challenge by four in ten of the finance leaders who responded to the survey. While finance chiefs are more likely to report that the organisation’s cost forecasts are generally reliable (39%), CEOs and MDs are less confident, with only 31% of them agreeing.
This dynamic is likely determined by each role’s proximity to forecasting. Finance leaders are more confident because of their familiarity with the process and direct data oversight, whereas CEOs and MDs are less exposed. Instead, they bear the consequences of forecast misses or last-minute budget changes.
Manual processes are still in effect
The data suggests businesses still rely on manual processes for forecasting tasks.
-
More than half (51%) of finance leaders say they use manual processes alongside other forecasting tools
-
A third (33%) report using them for a small part of the process
-
Some (12%) say they rely on them heavily for forecasting activities
-
Few (4%) claim they use no manual processes for forecasting
By and large, the reasons finance leaders give for still using manual processes relate to legacy technology and resource issues, including:
-
Data integration challenges (55%)
-
A need for flexibility or custom modelling (49%)
-
Limitations in forecasting tools (45%)
-
Cost or resource constraints (41%)
-
Historical processes that remain in place (41%)
While human judgement is critical to interpreting, reconciling, and validating financial data, over-reliance on manual data entry and disconnected systems can reduce the agility of the finance function.
For example, an average of 59% of finance leaders and CEOs report that manual data entry errors require employees to correct data before it’s used for decision-making. To stay competitive, organisations must implement AI in value-add areas that empower teams to focus on strategic analysis rather than routine administration.
So, where is AI used?
AI adoption in finance has transitioned from pilot phase to targeted, functional deployments. Today, just 3% of finance leaders who responded to the survey say they aren’t using AI in their analysis and forecasting tasks. Use cases include:
-
Revenue forecasting: 51% of finance leaders are using AI to estimate future income and improve the accuracy of financial projections
-
Risk analysis and demand forecasting: 45% and 43% of finance heads use AI for these respective processes, helping stress-test human assumptions
-
Scenario planning: 41% use AI to answer hypotheticals and support next steps
-
Real-time updates: 38% of finance leaders use AI to stay in the loop, suggesting teams may still struggle to connect AI to live data flows
-
Expense forecasting: 37% use AI for specialist predictions, reflecting cautious implementation in areas with tighter operational dependencies
-
Working capital forecasts: 37% of leaders integrate AI into their working capital forecasts, to keep an eye on their business’ liquidity
SAP Concur data helps debunk the myth of AI replacing real finance jobs. AI is being applied where it can improve productivity without forcing a major process change.
In 2026, finance leaders will need to strike a balance between human and AI touch – ensuring a competitive level of efficiency and accuracy from using technology and human expertise in tandem.
