The use of Artificial Intelligence Technologies in Corporate Reporting
7 minute read
Artificial Intelligence Technologies ("AI” or "AI technologies") such as Machine Learning (ML), Natural Language Processing (NLP) and Generative AI (GenAI) are developing rapidly and are being adopted across many sectors, processes and areas of business. In corporate reporting, however, the pace of adoption is more measured. This reflects the distinctive nature of the reporting environment and the demands placed on its outputs.
Our research
To understand how corporate reporting is adapting to the use of AI technologies, the FRC commissioned Lancaster University to undertake interviews and a survey with key stakeholders in the reporting process.
The resultant research, including the survey results, shows that the use of AI technologies is increasing, especially the use of GenAI, but adoption remains controlled, targeted and uneven. The direction of travel is clear, but integration is measured rather than disruptive. This reflects the design of the reporting environment, and the associated risk appetite of boards and organisations, rather than any lack of ambition.
The current role of technology in corporate reporting
Corporate reporting works within a framework defined by individual and organisational accountability. Financial statements, annual reports and investor communications are subject to detailed scrutiny often by auditors, investors and regulators. Therefore, the information produced must be reliable, consistent and capable of being defended and explained both internally and externally. It is this organisational context which constrains the rate at which technologies can be adopted.
However, our survey shows that mature technologies outside of AI technologies are already widely used in the corporate reporting process. Productivity tools such as spreadsheets and word processing are already near universal, with around 88% of respondents reporting active use. Similarly, core systems such as ERP (Enterprise Resource Planning) and general ledger platforms (68% adopted) and data visualisation and dashboard tools (66% adopted) are already widely embedded.
Adoption of AI technologies
The picture for AI technologies is more mixed. AI technologies such as machine learning and natural language processing are less popular than more mature technologies, with 18% and 8% of respondents adopting them respectively. Significant use cases for ML and NLP are ESG data collection and anomaly detection. Against this backdrop, the trajectory of GenAI is notable. 39% of organisations report current use, with a further 31% piloting and 18% considering adoption within a year, this indicates GenAI may be in line to be as core to the reporting process as mature tools like ERP within a short space of time. While our detailed analysis shows that the Financial Services sector is further ahead and that the most sophisticated use cases are typically limited to larger companies, relatively wide adoption across all types of companies points to a broader shift.
Adoption of AI Technologies chart
| Technologies | Adopted+Piloting (%) |
|---|---|
| ML | 42 |
| NLP | 41 |
| GenAI | 70 |
Definitions
Adoption of AI Technologies Groups chart
- ML: Machine Learning based Analytics
- NLP: Natural language processing for Narrative Analysis
- GenAI: GenAI Tools
This suggests that of the AI technologies, at least GenAI has moved beyond early experimentation and is now being evaluated more broadly across the reporting landscape (possibly reflecting the fact that tools like Copilot can be easier to implement rapidly across an entire organisation).
Our smaller group of interviewees broadly corroborated the survey findings with some who had already adopted AI characterising their use as ‘experimenting’ or ‘dabbling’; this is not surprising given the rapidly evolving nature of the tools and their capabilities.
Where AI is being used
From those participating in the interviews, it was clear that in their view corporate reporting has a low error threshold (numerically and narratively) because of its public facing nature. This is due to a concern that even minor inaccuracies can have significant reputational implications (for the individuals and/or the organisation). As a result, technologies that introduce this risk, such as GenAI, are approached cautiously by reporting teams.
The reporting process (and particularly the numerical elements) is one in which there already is a sophisticated use of tools such as Disclosure Management (tools that support the reporting process), ERP and data dashboard and visualisation tools. For interviewees this has meant that built-in AI functionality included within these mature tools is sometimes used automatically (while some turn off the functionality) with specific GenAI applications acting to support teams with more complex challenges as well as admin tasks.
For financial aspects of reporting, therefore, automation is most advanced in areas that are rule-based, data-intensive and lower risk. These include:
- Data extraction, consolidation and reconciliation
- Identification of anomalies and inconsistencies
- Preparation and validation of financial information
- Standardised compliance tasks
- Project management
In these areas, GenAI is acting to enhance already deployed tools and is used to improve efficiency and enhance review processes. It does “the heavy lifting” of processing large volumes of structured and unstructured data while operating within established processes and controls. For many this was identified as reducing the admin burden of the reporting process, a valuable outcome.
Interviewees also noted that GenAI is supporting Investor Relations (IR) teams in more investor-focused use cases such as analysis of earnings calls, investor questions and competitor announcements, distilling large amounts of unstructured information into key questions or themes to generate insights, including sentiment analysis which may feed into the annual report. IR teams also often use GenAI for tone and narrative consistency checks across materials to be published.
Where adoption is slower
By contrast, our interviews and survey present a more nuanced picture of adoption in areas requiring significant judgement or interpretation. This includes the drafting of management commentary, forward-looking statements and narrative explanations of performance.
What companies are currently using AI for in the reporting process chart
| Tasks | Companies (%) |
|---|---|
| Tone | 36 |
| X-Check | 44 |
| Trends | 56 |
| Copyediting | 57 |
| Drafting | 61 |
Definitions
What companies are currently using AI for in the reporting process chart
- Tone: Tone/sentiment checking
- X-Check: Cross-verify consistency
- Trends: Reporting trends / benchmarking
- Copyediting: Copyediting narrative
- Drafting: Draft narrative sections
Of survey respondents, 61% claimed they were using GenAI for drafting narrative, however, interviewees were more cautious. Interviewees indicated that where GenAI is used in narrative sections, it is typically focused on basic drafting tasks or for more compliance-driven content, often described as the ‘first draft’ activities, such as roll-forwards or describing charts and graphs, rather than higher-value and opinion-based commentary. The survey also indicated that 57% are using it for copyediting. This further corroborates that GenAI is being used as part of a human-led review and editorial workflow (co-worker role) rather than as an AI-only process (co-creator role).
This nuance in what is meant by drafting narrative may also help explain the split seen in the survey between senior management and more junior staff: with senior staff reporting that GenAI was used to support narrative drafting significantly less than more junior staff. Potentially, this reflects that first draft activities are often less visible to senior teams. This suggests there may be a disconnect developing between actual use within an organisation and management or board awareness of that use.
Across other areas the survey showed us that 36% of GenAI users are using it to interpret reporting standards and to track regulatory changes, and a further 39% and 41% respectively, are planning to do so in the next twelve months. Interviews also illustrated using GenAI as a checklist. For example, one interviewee discussed piloting AI to identify areas of improvement in the annual report versus areas highlighted as key issues by the FRC, and another explained how they are using Copilot to evaluate draft disclosures by benchmarking against multiple auditor reporting checklists.
Opportunities
In the interviews the primary benefit of AI technologies identified by respondents was improved efficiency by reducing the time required for routine and repetitive tasks and allowing reporting teams to focus on higher-value activities, including benchmarking and review.
Importantly, interviewees suggested that these efficiency gains are not typically being realised as simple cost savings. Instead, they are being reinvested in the reporting process. At present, however, these more advanced use cases are concentrated among larger and more resourced organisations, with smaller entities focused more on basic tasks.
Risks
We also heard about the risks. The cautious approach observed in the use of AI in corporate reporting is underpinned by a number of well-recognised risks.
These include:
- Risk of error: Even small inaccuracies can have significant consequences, reinforcing the need for robust validation and review processes. Furthermore, GenAI tools may still be prone to hallucinations and could misinterpret standards or regulation, reinforcing the need for expert human review.
- Data quality: AI outputs are dependent on the quality of underlying data, which in many organisations remains fragmented or inconsistent.
- Transparency and explainability: Some AI systems operate in ways that are not easily understood or explained, which can limit their acceptability in reporting contexts.
In addition, interview participants highlighted uncertainty regarding acceptable use of AI as a key barrier. In particular, there is sometimes a lack of clarity within organisations on how far AI can be used in externally reported content and what level of disclosure or oversight is expected. Reporting teams are reluctant to use GenAI to produce more complex narrative content, with a concern not only to the risk of error, but also the importance of maintaining an authentic management voice with investors.
This distinction between supporting basic narrative production and generating higher value narrative content is emerging as a key boundary in current practice.
Another barrier to adoption of AI identified by participants is the lack of appropriate skills and related resourcing gaps. This will require more robust training, with current training being patchy according to interviewees and primarily focused on risk mitigation. These findings were also mirrored in the survey.
Additionally, whilst 91% of respondents who said they use GenAI are using Enterprise GenAI tools, 24% said they are also using public AI tools – this could heighten cyber and data risks.
Governance and oversight
A consistent theme across the research (both interviews and surveys) was the view among participants that successful integration of artificial intelligence depends on effective governance.
Key elements identified by many as already in place included:
- Clear policies on acceptable use
- Defined accountability for outputs
- Robust eview and validation processes
- Appropriate levels of human oversight
- Engagement with internal auditors and other assurance providers
Our survey noted that whilst many companies have taken steps in these areas, there remains variation in practice. In particular, there is evidence of gaps in human oversight (with only 44% of companies having mandated this) in spite of "human in the loop” review often being mentioned as essential.
Conclusion
This research provides a useful snapshot of the current stage of AI technology adoption in corporate reporting. Adoption is increasing but remains shaped by the characteristics of the reporting environment, the need for accuracy, the importance of trust, and the requirement for accountability.
This results in an approach in which:
- AI technology is widely used for process-intensive and lower-risk tasks, increasing efficiency and reducing burden.
- Judgement and high value narrative remain an overtly human process (although GenAI is supporting at the periphery).
- Governance and oversight are critical enablers of adoption.
The opportunity for all is now to support this transition ensuring that the benefits of artificial intelligence can be realised, while maintaining the quality and integrity of corporate reporting.
Over the next few months, the FRC will publish a series of case studies highlighting how AI is being used by companies in their corporate reporting process. The FRC will continue to monitor the developing use of AI technologies and will support and engage where the market needs clarity.
Background
The FRC commissioned Lancaster University (with a team including academics from Loughborough University) to undertake research on our behalf. Findings and conclusions are derived from a combination of 39 interviews with senior members of reporting and investor relations teams (plus other external corporate reporting stakeholders including advisors, third party software vendors and auditors) and results from an online survey that generated 103 responses. The survey was distributed using a multi-channel strategy including a commercial research panel provider, email invitations sent to the IR lead at FTSE350 companies, and completion requests distributed by the FRC and professional bodies and representative groups.
The survey went live on 17 February 2026 and closed on 21 April 2026.