12 Building Blocks for the Controlled Use of Audit Tooling

12 Building Blocks for the Controlled Use of Audit Tooling

Building Blocks: • (Gen)AI tools are used safely and in a controlled manner. • The results from the (Gen)AI tool are verifiable or reproducible. • The external auditor bears final responsibility for the results from the (Gen)AI tool.

Building Blocks: • Use an appropriate framework for information security. • Formalize policy regarding audit tooling. • Monitor the effect of audit tooling on the quality of the statutory audit.

Building Blocks: • Perform the appropriate IT activities that oversee the reliability and usability of the input data. • Establish that relevant and reliable input data are used to obtain sufficient and appropriate audit information. • Ensure a consistent ETL process and document it.

Building Blocks: • Develop knowledge and skills in line with innovation. • Integrate the application of audit tooling into the audit strategy and audit plan. • Follow up on the results of audit tooling in the correct manner.

Risk Management The accounting firm has its risk management and control in order. This is the foundation. 1

Input Data The input data are relevant and reliable for obtaining sufficient and appropriate audit information. 2

Audit Tooling The audit tool is implemented in a controlled manner. 3

(Gen)AI Tooling The (Gen)AI tool is applied in a controlled manner. 4

Ensure a solid foundation !

SUPERVISION REPORT DECEMBER 2025 12 Building Blocks for the Controlled Use of Audit Tooling

In Short Advanced audit tools – including (Gen)AI – offer accounting firms ample opportunities: more efficient processes, higher quality of audit activities, and more attractive work. The AFM encourages innovation, provided it is responsible and controlled. Some larger accounting firms are leading in implementing policy for tooling in statutory audits. For others, the foundation is less stable. There is room for improvement everywhere. Therefore, the AFM offers 12 building blocks to evaluate what is already going well and where strengthening is needed. The bottom three layers of the tower must first be in good order before deploying innovative technology such as (Gen)AI. Check the foundation, strengthen where necessary, and continue building on quality and trust.

SUPERVISION REPORT Contents Introduction 3

1. Prerequisite 1: The accounting firm has its risk management and control in order 12 1.1 Building Block 1: Use an appropriate framework for information security............................................... 12 1.2 Building Block 2: Formalize policy regarding audit tooling ............................................................................... 14 1.3 Building Block 3: Monitor the effect of audit tooling on the quality of the statutory audit .......... 16
2. Prerequisite 2: The input data are relevant and reliable for obtaining sufficient and appropriate audit information 18 2.1 Building Block 4: Perform the appropriate IT activities that oversee the reliability and usability of the input data ................................................................................................................... 18 2.2 Building Block 5: Establish that relevant and reliable input data are used to obtain sufficient and appropriate audit information .....................................................................................................20 2.3 Building Block 6: Ensure a consistent ETL process and document it...........................................................22
3. Prerequisite 3: The audit tool is implemented in a controlled manner 24 3.1 Building Block 7: Develop knowledge and skills in line with innovation........................................................24 3.2 Building Block 8: Integrate the application of audit tooling into the audit strategy and the audit plan......................................................................................................................................................26 3.3 Building Block 9: Follow up on the results of audit tooling in the correct manner...........................................27
4. Prerequisite 4: The (Gen)AI tool is applied in a controlled manner 31 4.1 Building Block 10: (Gen)AI tools are used safely and in a controlled manner............................................................... 31 4.2 Building Block 11: The results from the (Gen)AI tool are verifiable or reproducible .....................33 4.3 Building Block 12: The external auditor bears final responsibility for the results from the (Gen)AI tool ..................................................................................................................................................35

12 Building Blocks for the Controlled Use of Audit Tooling 3 SUPERVISION REPORT Introduction 1 Committee of European Auditing Oversight Bodies (“CEAOB”) - Challenges and applications of advanced technologies in audit firms (published in October 2024) 2 AFM – Trend View 2026 (published in November 2025) 3 International Forum of Independent Audit Regulators (“IFIAR”) - Use of technology in audits – observations, risks and further evolution (published in March 2025) 4 In accordance with NBA Guideline 1141, we use the following definition for data analysis: “Data analysis is the discovery of patterns, deviations, inconsistencies, and the extraction of further useful information about the object of the investigation through analysis, modeling and visualization with a view to planning or executing the assignment.” 5 By (Gen)AI, an ‘AI model for general purposes’ is meant as defined in Article 3 definition 63 of Regulation (EU) 2024/1689 of the European Parliament and the Council. 6 These accounting firms were selected from the data provided by RV-accounting firms to the AFM regarding 2024. This data is presented in Figure 4.3 on page 43 of Trend View 2026. Technological developments and digitalization are rapidly changing the way accounting firms perform their activities, strengthening the importance of controlled use of audit tooling. These developments offer opportunities, such as increased efficiency and quality of audit work and more attractive work.1 At the same time, they bring risks: inadequate control of the technology or incorrect application can undermine the quality of statutory audits and lead to operational vulnerabilities.2, 3

We identify several ‘drivers’ behind the use of new audit tooling and the increased attention for information security at accounting firms. Examples include the availability of new technologies, efforts to address staff shortages, and pressure from private-equity parties focusing on efficiency. These trends make audit tooling an increasingly integral part of the audit practice. At the same time, risks in the area of information security are increasing. This is partly due to the increasing risk of cybercrime: accounting firms process confidential data and are therefore attractive targets. Additionally, new legislation – such as the Digital Operational Resilience Act (DORA), the Baseline Information Security Government version 2 (BIO2), and the Network and Information Security Directive 2 (NIS2) – is generating more attention regarding digital resilience.

We welcome innovation within the sector and emphasize the necessity of controlled use of new technology. Accounting firms possess large amounts of data from parties they audit. It is crucial that this data is processed, transferred, and stored securely, and that the data serving as input for audit tools is relevant and reliable to obtain sufficient and appropriate audit information to support the external auditor’s opinion. Furthermore, audit tools must perform reliable and transparent analyses to ensure the quality of statutory audits. This is essential for trust in the accounting sector.

Therefore, we conducted an exploratory study into the use and scope of audit tooling in statutory audits and its impact on the quality of these audits. This concerns audit tooling in the broadest sense: from dossier software to data analysis tools4 and (Gen)AI5, in all phases of the audit process. An overview of tools we encountered during our exploration is presented on page 30 of this report. The study was conducted at eleven accounting firms with a regular license6 (“RV-accounting firms”) and at two accounting firms with a license for statutory audits at entities of public interest (“OOB-accounting firms”). During the on-site investigations, we assessed components of the quality management system that oversee the use and implementation of audit tools. We also explored the use and scope of audit tools in two statutory audits to gain insight into their influence on audit quality. Additionally, we held in-depth discussions with various stakeholders, including suppliers of eight

12 Building Blocks for the Controlled Use of Audit Tooling 4 SUPERVISION REPORT different audit tools and various industry and professional organizations. Finally, in addition to the existing data for RV-accounting firms, we conducted a data request with the six OOB-accounting firms to map the organization-wide use of audit tooling within this population.

There is still much to gain in terms of governance and control of audit tools within the sector. Although some larger accounting firms appear to be leading regarding information security policy and tool-specific policy at the organization level, the majority of organizations have less formalized frameworks. Our research shows, among other things, that more structural attention is needed in the area of risk control and the correct application of audit tooling. This applies in particular to the group of accounting firms with limited policy, but also for organizations that have further developed policy frameworks, there is room for improvement.

To help the sector move forward, our report contains important prerequisites for the controlled use of audit tooling. The prerequisites are structured into 12 building blocks. With these building blocks, accounting firms can (further) work on a robust structure for the controlled implementation of audit tooling. 7 AFM – Trend View 2026: Given the risks identified by the AFM, this topic will receive increased attention in (testing) investigations in the coming years. (published in November 2025)

What does the AFM expect? A structure that does not eventually wobble or even collapse because one or more prerequisites are lacking. The AFM encourages accounting firms to critically assess how they use audit tooling, so that it is implemented in a controlled manner in the statutory audit. Answer at least the following three questions for your organization: Which building blocks must or can the organization use? Does the organization need all building blocks for a robust structure? And: Are there possibly other relevant building blocks besides the 12 included in this report? It is up to accounting firms to answer these questions, depending on factors such as the nature of the audit and the audit tooling used. Organizations must make well-considered choices in this and comply with applicable legislation and regulations.

Our call to the sector: use the building blocks in this report to evaluate what is already going well and where improvement is needed. Check the foundation, strengthen where necessary, and continue building on quality and trust.

The AFM is expected to devote more attention in the coming years to the controlled use of audit tooling in further (testing) investigations.7

12 Building Blocks for the Controlled Use of Audit Tooling – from 4 Prerequisites Prerequisite 1: The accounting firm has its risk management and control in order Why this prerequisite? Good risk management and effective risk control are essential for the responsible use of audit tooling. The management board of the accounting firm plays a crucial role in this. The board is responsible for setting up a robust framework for information security and developing policy that facilitates controlled application of audit tooling. Additionally, it must ensure the effective rollout of audit tools8, tool-specific policy, and the setup of monitoring activities that make the use and scope of audit tools in the audit practice measurable.

What does the AFM see in practice regarding this prerequisite? Our research shows that risk management and control around audit tooling are not always adequately arranged. We see shortcomings in the information security framework and insufficient grip on third parties at various accounting firms, resulting in cyber risks not being sufficiently controlled. Furthermore, tool-specific policy for the use and monitoring of audit tools is often not formalized, posing risks for the execution of the statutory audit. Directors are not always aware of these risks. 8 Financial Reporting Council (“FRC”) - Certification of Automated Tools and Techniques (published in June 2025)

Building Block 1: Use an appropriate framework for information security. Information security is essential, especially as the audit process becomes increasingly data-driven. Nevertheless, we see that many accounting firms are vulnerable due to ad-hoc policy, limited grip on third parties, and insufficient awareness of risks among policy makers. A robust risk management framework provides guidance to design information security in a future-proof manner and prevent the “hacker as a wake-up call.”

Building Block 2: Formalize policy regarding audit tooling. Accounting firms are increasingly using audit tooling as part of their audit approach. Nevertheless, many organizations lack formalized policy on the use of audit tooling. As a result, the application is not always effective or efficient for the audit objective or the overall audit approach. By formalizing policy, the board gains grip on the use of audit tooling within the accounting firm.

Building Block 3: Monitor the effect of audit tooling on the quality of the statutory audit. Audit tools often offer functionality for usage monitoring, making the use and scope measurable. By systematically monitoring, the board gains insight into the effect of audit tooling on the statutory audit. For example, by assessing whether the correct functionalities are used for audit activities, or if one falls back on traditional methods. This allows for timely adjustments and the stimulation of desired behavior. The AFM observes that monitoring functionalities of audit tools are often not used, resulting in incorrect use, technical errors, or professional shortcomings going unnoticed.

12 Building Blocks for the Controlled Use of Audit Tooling 6 SUPERVISION REPORT Prerequisite 2: The input data for the tool are relevant and reliable for obtaining sufficient and appropriate audit information Why this prerequisite? Audit tools can help in collecting, sorting, filtering, and analyzing data from the audit client. The input data in audit tooling must then be relevant and reliable. The quality of the results is directly dependent on the relevance and reliability of this data. Therefore, it is necessary for the external auditor to perform appropriate activities to establish that the input data are relevant and reliable, including accuracy and completeness as needed for supporting the opinion.

What does the AFM see in practice regarding this prerequisite? We note that external auditors do not always perform the appropriate activities to establish the relevance and reliability of input data (including the completeness and accuracy of non-financial data elements). For example, insufficient attention is regularly paid to ineffective General IT Controls (“GITC’s”)9 or the source verification of data obtained by the auditor. We see that data from systems with ineffective internal controls are used without taking additional safeguards. Also, a (policy-embedded) standardized process to ensure that relevant and reliable (complete and accurate) data from the audit client have been received is often missing.

Building Block 4: Perform the appropriate IT activities that oversee the reliability and usability of the input data. The external auditor must have insight into the influence of the audit client’s GITC’s on the reliability and usability of the data. This requires the external auditor to perform sufficient in-depth activities to gain insight into the risks of unreliable or unusable data. Our research indicates that auditors do not always give sufficient attention to GITC’s and other IT-related risks to secure the conclusion on the reliability and usability of the input data. 9 NV COS 315.12d: This paragraph gives the definition of GITC’s used in this report.

Building Block 5: Establish that relevant and reliable input data are used to obtain sufficient and appropriate audit information. When the external auditor receives data from the audit client and uses it as input data for audit tooling, it is important that the appropriate activities are performed to establish that relevant and reliable data are used to obtain sufficient and appropriate audit information. In practice, we see that the external auditor does not always take the appropriate safeguards to ensure that the relevance and reliability of the received data are secured with regard to the audit objective. For example, by establishing the completeness and accuracy of the received data by linking them at the line level to underlying administrative records, and where relevant, to original source documents.

Building Block 6: Ensure a consistent ETL process and document it. Accounting firms must ensure a consistent Extract-Transform-Load (ETL) process, preferably documented in policy. The external auditor must document the ETL process with sufficient detail in the audit file. A qualitative ETL process makes activities traceable, even when using specialists or external parties. A carefully arranged ETL process within the organization also contributes to the consistent evaluation of the usability, reliability, and relevance of data used in the audit, which can positively influence audit quality. In practice, we see good examples, but also room for improvement.

12 Building Blocks for the Controlled Use of Audit Tooling 7 SUPERVISION REPORT Prerequisite 3: The audit tooling is implemented in a controlled manner Why this prerequisite? Audit tooling is implemented as part of the audit process, but only with correct use does it contribute to the quality of the audit. For controlled application, it is important that the engagement team possesses the right knowledge and skills, that it is clear what role the audit tool plays in the overall audit plan, and that the results of audit tooling are handled in the correct manner.

What does the AFM see in practice regarding this prerequisite? In practice, we see that audit tooling is not always applied correctly. Controlled application requires that the external auditor understands the functionalities of the audit tooling, how they are used for a qualitative audit, and how the results contribute to obtaining sufficient and appropriate audit information. Here, we see that sometimes the audit objective is overlooked, and the application of the tool does not fully align with the objectives in the audit plan.

Building Block 7: Develop knowledge and skills in line with innovation. Accounting firms often have a wide range of audit tools with varying functionalities. Nevertheless, the full potential is not always utilized. This may be because engagement teams are not aware of all possibilities or do not possess the right knowledge to apply the tooling well. When knowledge or skills are lacking, accounting firms can organize training programs and provide practical guidance to employees, or ensure that the right knowledge and competencies are brought in-house from other specialties.

Building Block 8: Integrate the application of audit tooling into the audit strategy and the audit plan. Audit tooling plays a supporting role in the audit process; it is not an end in itself. In practice, we see that audit tooling is not always used based on assessed audit risks or the overall audit plan, resulting in the outcomes (in some cases) providing weaker audit evidence. It is important that audit tooling is integrated into the audit approach, aligned with identified risks and the amount of audit evidence that can be obtained with the tooling.

Building Block 9: Follow up on the results of audit tooling in the correct manner. Audit tooling is used in statutory audits to generate audit evidence or other outcomes. We see room for improvement in how these outcomes are handled, for example, with (potentially significant) exceptions and deviations that can lead to audit differences or other findings. It is important that the user of an audit tool handles the outcomes correctly to secure audit quality.

Prerequisite 4: The (Gen)AI tool is applied in a controlled manner Why this prerequisite? (Gen)AI brings new, inherent risk factors due to the nature and complexity of the technology. The previous 9 building blocks are self-evidently also relevant for the use of (Gen)AI tooling, but controlled application of (Gen)AI tooling requires 3 additional building blocks.

What does the AFM see in practice regarding this prerequisite? With (Gen)AI tooling, we more often see a risk of overreliance than with traditional audit tools, due to the form of the presented outcomes. Outcomes of (Gen)AI tools can appear convincing due to the anthropomorphic characteristics of some tools. These tools mimic human characteristics, causing the user to often trust the outcomes more quickly and evaluate them less critically. That the results of (Gen)AI tooling can appear convincing due to the structure and/or form can cause the ‘halo effect’. Furthermore, it is not always transparent how outcomes are generated – this is due to the st

12 Building Blocks for the Controlled Use of Audit Tooling 8 SUPERVISION REPORT structure of the models and the lack of explainability. This opacity increases the risk that users accept outcomes without sufficient critical scrutiny.

Building Block 10: (Gen)AI tools are used safely and in a controlled manner. The use of (Gen)AI tools requires specific safeguards to prevent misuse, data leakage, or biased outcomes. Organizations must ensure that (Gen)AI tools are selected, configured, and monitored in a way that aligns with security standards and regulatory requirements. This includes regular testing for bias, accuracy, and robustness against adversarial attacks.

Building Block 11: The results from the (Gen)AI tool are verifiable or reproducible. Due to the probabilistic nature of (Gen)AI, outcomes may vary. It is essential that the processes surrounding (Gen)AI allow for the verification of results. This means that inputs, parameters, and model versions are documented so that outcomes can be reproduced and audited. Without verifiability, the audit evidence provided by (Gen)AI lacks the necessary reliability for statutory purposes.

Building Block 12: The external auditor bears final responsibility for the results from the (Gen)AI tool. The use of (Gen)AI does not absolve the external auditor of professional responsibility. The auditor must critically evaluate the outputs of (Gen)AI tools, understand their limitations, and ensure that the final audit opinion is based on sufficient and appropriate audit evidence. The auditor must be able to explain and justify the role and impact of (Gen)AI in the audit process to regulators and stakeholders.