From Numbers and Charts to Insights and Explanations
Financial reporting organizes recorded data according to agreed-upon definitions. General ledger accounts are consolidated into reporting lines, amounts are allocated to periods and entities, and results are compared with, for example, the budget, forecast, or the previous year. Tables and charts visualize these results.
The next step is to determine which trends are relevant and what might have caused any variances. For example, a revenue variance could be related to volume, price, product mix, currency, timing, or changes in the organizational structure. Internal reports therefore often include a note describing the key results, variances, and possible explanations.
The meaning of a figure is determined by its context. The period, currency, entity, accounting principles, dimensions, and basis of comparison determine exactly what a result represents. Without this context, the same amount or percentage can mean something entirely different.
For recurring reports, traditional automation can reliably perform the calculations. Report definitions determine which accounts, dimensions, periods, and calculations are included in a report. Consolidation, elimination, and currency rules process the figures in a reproducible manner. The more challenging step is analyzing the results and selecting and formulating a relevant explanation.
Where Traditional Automation Falls Short
Fixed reports and dashboards display predetermined figures, comparisons, and charts. Rules can flag variances above a set amount or percentage. This works well when it is clear in advance which comparison is relevant and when a variance warrants attention.
However, the significance of a deviation depends on the context. A small percentage deviation on a major cost item may be more significant than a sharp increase within a one-time, limited category. A minor deviation may also be part of a longer-term pattern that only becomes apparent when multiple periods, entities, or dimensions are viewed together.
The explanation requires even more context. A difference in revenue may be related to volume, price, product mix, currency, timing, or a change within the organization. Describing all these combinations using traditional logic requires many separate rules and text templates. As a result, the explanation quickly becomes general or mechanical, while employees still have to search for the relevant information in tables, charts, and underlying transactions.
Furthermore, a calculated correlation does not constitute a proven cause. The fact that revenue and sales volume are rising simultaneously makes a correlation plausible, but does not rule out other explanations. An automatically generated explanation should therefore not express more certainty than the available data justifies.
How AI Can Accelerate Analysis
AI can analyze financial results across multiple periods and dimensions and highlight relevant deviations, trends, and patterns. In doing so, it can look beyond fixed thresholds. The absolute impact, recent developments, underlying transactions, and deviations from the usual pattern can also be taken into account.
In this way, AI can identify which developments—in areas such as revenue, margin, costs, working capital, or cash flow—are likely to require further explanation. The analysis can then zoom in on underlying factors, such as volume, price, product mix, currency, timing, entity, department, or product group.
AI can also establish connections between different tables, charts, and detailed data. As a result, a deviation at the aggregate level can be traced back to the periods, business units, or transactions that contribute most significantly to it. This speeds up the analytical process, but does not automatically mean that a detected correlation constitutes a proven cause.
Generative AI can formulate a draft explanation based on the selected results. This explains what has changed, where the most significant variances are, and which factors may be contributing to them. Every value and explanation mentioned must remain traceable to the figures, charts, and underlying source data used.
The financial figures themselves are not calculated by a language model. The existing reporting and consolidation logic remains responsible for amounts, subtotals, and comparisons. AI analyzes these audited results and supports the selection and formulation of relevant insights.
A distinction must continue to be made between established facts, calculated correlations, and possible explanations. A change in price, volume, or product mix can be mathematically substantiated. An assumed commercial or operational cause may still need to be confirmed by a responsible employee.
The result is a faster path from financial figures and charts to actionable management information. AI helps identify relevant developments, examine correlations, and draft an initial explanation, after which the responsible employee reviews and supplements it.
AI quickly analyzes financial results, identifies relevant deviations and trends, and generates a well-founded, traceable explanation to accompany the figures and charts.