Reconcilingdata, identifying discrepancies, preparing management commentary: certain steps in the reporting process can now be supported by artificial intelligence. But producing an analysis more quickly is not enough to inform a decision. For the management controller, the transformation is as much about the tools as it is about the ability to verify the numbers, discuss assumptions, and support decision-making.
From Traditional Reporting to AI-Powered Management
The role of a management controlleris built around one essential mission: to help an organization understand its results and use its resources in alignment with its objectives. Budget preparation, cost analysis, margin monitoring, forecasting, and communication with operational managers form the foundation of this role. Reporting—that is, the regular production of information for management purposes—is only one component of the role. Interpreting results and proposing corrective measures are already part of the job, as noted in the APEC job description1.
Artificial intelligence, therefore, does not create the analytical aspect of management accounting. Rather, it introduces new possibilities in data preparation, variance analysis, and the drafting of reports that support this function. The question then becomes less about the number of reports produced and more about their reliability and usefulness to decision-makers.
Significant challenges remain in accessing actionable information. An international survey published in May 2025 by Deloitte and the Institute of Management Accountants, involving more than 440 finance and accounting executives, indicates that 54% of respondentsdo not have the necessary reporting in place or believe they need to improve the transparency of their cost and profitability information. This finding reflects a need expressed by professionals, not proof that AI alone will resolve it2.
It is in this area that the transformation of the profession is taking place: automating what can be automated, while strengthening the ability to understand what the numbers actually tell us. A generated comment is not necessarily a proven explanation, and a recommendation does not constitute a decision.
How AI Is Transforming Management Control Practices
It is essential to make a distinction up front. Adding up journal entries, updating a dashboard, or calculating a budget variance does not necessarily require artificial intelligence. These tasks can be based on standard calculation rules.
AI’s specific contribution lies, in particular, in recognizing unusual patterns, assisting with certain analyses, and generating text based on provided information.Generative AIthus produces content, but the fact that it is convincingly worded does not guarantee its accuracy—a limitation highlighted by the CNIL3.
Several documented instances illustrate this development.
Support data reconciliation prior to reporting
Microsoft’s Financial Reconciliation tool identifies matches between two data sets in Excel, then distinguishes between reconciled transactions, potential matches, and discrepancies. Users can modify the suggested rules and locate the source transactions. For management accounting, one possible use is to prepare a reconciliation between an accounting statement and a tracking table. The tool facilitates the review but does not eliminate the need to verify that the periods and scopes being compared are identical[4].
Prepare comments on discrepancies
Microsoft’s Variance Analysis feature describes the process of identifying variances between budgeted, forecasted, and actual data, followed by the generation of an editable comment. In particular, it allows users to focus the analysis on specific products, regions, or departments. However, the documentation accessed in September 2026 is still marked as a pre-release version. More importantly, identifying the elements that mathematically contribute to a variance does not reveal the business or organizational causes behind it[5].
Identify anomalies and forecasting biases
The Oracle IPM Insights documentation describes analyses that can flag outliers, missing data, or a recurring tendency to overestimate or underestimate results. It also allows for a comparison between teams’ forecasts and those generated by the models. An anomaly then becomes a point to investigate, not an error to be automatically corrected: it may correspond to an actual business event[6].
These features demonstrate that automation can go beyond simply formatting reports. However, they do not, on their own, provide independent evidence of productivity gains or improved decision-making across all companies.
For a deployment, the relevant question therefore remains a practical one: How much time is actually saved once data preparation, result verification, and comment correction are factored in?
A New Role for the Management Controller
The transformation does not involve turning the management controller into a passive overseer of algorithms. On the contrary, it can reinforce a responsibility that is already central to the role: giving meaning to results and making assumptions open to discussion. Collaborating with operational staff and formulating proposals for improvement are already part of the role’s scope1.
Define what the indicators actually measure
Before asking a system to explain a trend, it is necessary to specify the rules used. Does the margin include logistics costs? Are the figures compared on a like-for-like basis? Has an organizational change shifted costs between two departments? Without these clarifications, an analysis may be numerically consistent but economically misleading.
Compare the explanations with operational realities
A decline in profitability may result from a sales discount, a change in the products sold, a rise in costs, or a deliberate investment in quality. The controller’s role is to verify the assumptions with the relevant teams, rather than simply accepting the most convenient explanation offered by an assistant.
Distinguishing Between Objectives, Forecasts, and Scenarios
A goal expresses an ambition. A forecast estimates a likely outcome based on available information. A scenario explores the consequences of specific assumptions. Confusing them is equivalent, for example, to presenting a target budget as a proven trajectory.
Guiding decision-making without taking the place of decision-makers
The management controller can compare several options, break down their costs, and identify uncertainties. The decision to invest, change a price, or reorient a business activity then rests with the appropriate managers. Automating reporting does not eliminate this division of responsibilities.
The value of a professional thus lies in the soundness of the reasoning that links the data, its interpretation, and the options under discussion.
What Skills Does a Management Controller Need in the Age of Generative AI?
Financial expertise remains the foundation of the profession. Understanding cost structure, profitability mechanisms, and the effects of changes in assumptions remains essential for evaluating automatically generated results.
Nevertheless, professionals have identified a need for training. A CIMA survey published in December 2025, which polled 1,446 executives and finance and accounting managers, indicates that 56% of respondentsconsider generative AI to be their primary skills gap. This survey covers finance and accounting functions in the broadest sense, not just French management accountants7.
Technical and digital skills
- Technical and digital skills tailored to specific uses. Auditors must be able to identify the source of data, how it has been processed, and its limitations. Understanding the connections between accounting, decision-making tools, and management reports is essential for verifying analyses. This does not necessarily mean becoming a data engineer, but rather being able to communicate with data specialists and articulate specific requirements.
Analytical and decision-making skills
- Analytical reasoning capable of withstanding plausible answers. Using an assistant requires knowing how to ask it for its sources, verifying its calculations, and distinguishing between an observation and a hypothesis. Mastering the instructions given to the model is not enough: the professional must also be able to identify what is missing from the answer and what warrants further investigation.
Ethical, Legal, and Organizational Skills
- An understanding of ethical and legal requirements. A report may contain information about employee compensation or identifiable customers. This personal data must be processed in accordance with the GDPR, particularly with regard to purpose, data minimization, retention periods, and access security. Confidential business information also requires careful consideration of the tools used and their contractual terms. The CNIL recommends establishing guidelines for the use of generative AI and training users, rather than allowing individuals to freely transmit business documents to an external service[8].
- Interpersonal skills to facilitate decision-making. Explaining uncertainty, challenging an assumption without stalling the discussion, and presenting several options remain essential. A technically correct result only informs a decision if it is understood by those who must act on it.
A useful training program would therefore benefit from drawing on real-world situations: examining an erroneous comment, tracing the source of a figure, comparing two forecasts, and justifying a recommendation. The challenge is not only to learn how to use AI, but also to learn how to manage its contributions.
Can artificial intelligence make performance management more reliable?
It can contribute to this, but reliability must be assessed on several levels: the accuracy of the figures, the relevance of their interpretation, and the robustness of the forecasts. An improvement on one level does not guarantee improvements on the other two.
What the evidence allows us to conclude
A fictional educational example helps illustrate this distinction.
A company plans to sell 1,000 units at 100 euros each, for a total revenue of 100,000 euros. It ultimately sells 1,200 units at 90 euros each, for a total of 108,000 euros. Revenue therefore exceeds the budget by 8 percent.
With a constant variable cost per unit of 60 euros, the contribution margin nevertheless fell from the projected 40,000 euros to the actual 36,000 euros, a decrease of 10 percent. In this case, the contribution margin is calculated as revenue minus variable costs, without taking fixed costs into account.
A comment limited to “volume growth supports strong performance” would be incomplete. The calculations—which can be performed without AI—show that the increase in revenue is accompanied by a decline in this margin. However, they do not allow us to conclude that the sales strategy is necessarily flawed: the company may have accepted lower immediate profitability to achieve an explicit objective.
The auditor must then verify this objective, its expected effects, and the conditions under which the decision will be reassessed. The figure describes a consequence; it does not, on its own, determine the appropriateness of the decision.
The same caution applies to forecasts. A model that replicates past data very well may perform less well over new time periods. The evaluation must use data that were not used to train the model and compare its results to a benchmark method, over a relevant forecast horizon. The methodological work of Hyndman and Athanasopoulos specifically highlights this distinction between the quality of the fit and the actual accuracy of forecasts9.
Limits to Watch For
Several risks also require special attention. A generative model may produce a plausible but unsubstantiated justification. An analysis may overlook information that is missing from the provided data set. Excessive reliance on the response may lead to a reduction in necessary verification, as the CNIL points out3.
To evaluate a system, it is therefore best to use concrete criteria: the rate of incorrect reconciliations, the number of corrected comments, the traceability of figures, the accuracy of forecasts, and the total time spent on verification. The goal is not to produce more analyses, but to obtain usable analyses without creating a disproportionate verification burden.
What will the role of a management controller look like in the future with AI?
The most plausible scenario is not one in which management control is uniformly autonomous. It will depend on the quality of the data, the organization of work, and the decisions the company is willing to delegate to the tools.
- More automated case preparation: systems will be able to aggregate information, flag inconsistencies, and prepare summaries. The time saved will only translate into better advice if the organization actually reallocates that time to analysis and communication with the client.
- More continuous monitoring of risks: fraud, transaction anomalies, and changes in certain indicators can be tracked more frequently. This monitoring also increases the volume of alerts and therefore requires prioritization and control mechanisms.
- Generative assistants will play a greater role in customer service: they will be able to facilitate document searches or draft responses. Their deployment will require well-managed knowledge bases, confidentiality controls, and a precise definition of the situations in which a human must take over.
- Agent-based systems in controlled environments: some assistants will be able to perform multiple tasks in sequence. In sensitive processes, the key issue will be the permissions granted to the system: reading, suggesting, transmitting, modifying, or executing are not equivalent levels of autonomy.
- More structured AI governance: The ECB has included AI strategy, governance, and risk management among its supervisory priorities. In particular, it plans to continue targeted monitoring of banks’ generative AI applications.8
- Increased regulatory compliance requirements: Bankers will have to operate in an environment where banking regulations, data protection, digital resilience, and the AI Act all intersect. Not all AI systems will be subject to the same requirements; the use case, purpose, and level of impact will remain key factors.
The time freed up will therefore only become time for analysis if the company organizes this reallocation. The transformation of the profession remains as much a matter of management and training as it is a technological issue.
AI-Enhanced Management Control for Informed Decision-Making
Automating reporting does not change the purpose of management control: to make performance understandable and help the organization make informed decisions. It does, however, shift some of the focus toward data validation, discussing assumptions, and justifying recommendations.
This responsibility goes beyond simply seeking immediate cost reductions. A trade-off may improve a financial metric while undermining the company’s quality, timelines, skills, or future capabilities. The controller’s role is also to highlight these effects and explain what the dashboard does not measure.
The value of a tool must therefore be assessed based on the problem it solves, its total cost, and the resources it requires. Conventional automation may be sufficient when a calculation rule meets the need. A more complex model is relevant only if its added value is demonstrated.
Learn more
To further explore these topics— the reliability of analyses, the oversight of tools, and the role of professional judgment— check out four related articles on the aivancity blog.
References
[1] Apec. (n.d.). Management Controller (M/F). View source
[2] Institute of Management Accountants and Deloitte. (May 22, 2025). New Deloitte, IMA Survey Highlights Emerging Technology Trends in Cost Accounting and Profitability Reporting. View source
[3] CNIL. (July 18, 2024). CNIL Q&A on the use of a generative AI system. View source
[4] Microsoft. (2025). Reconcile data with the Financial Reconciliation agent. Microsoft Learn. View the documentation
[5] Microsoft. (2026). Analyze variances with the Variance Analysis feature. Microsoft Learn. Preview documentation, accessed September 24, 2026. View the documentation
[6] Oracle. (n.d.). About IPM Insights. Oracle Cloud Documentation, accessed September 24, 2026. View the documentation
[7] AICPA & CIMA. (December 17, 2025). Finance professionals say their organizations must better prepare for anticipated AI disruption. View source
[8] CNIL. (n.d.). The Six Key Principles of the GDPR. View source
[9] Hyndman, R. J., and Athanasopoulos, G. (2021). Forecasting: Principles and Practice, 3rd edition. Section Evaluating point forecast accuracy. OTexts. View the book