Articles

What is the role of AI in corporate reporting? That’s the question the Financial Reporting Council’s latest research set out to explore. The findings show that while AI is now embedded in many listed companies’ approach corporate reporting — this is perhaps not in the ways you might expect.

Lancaster University was commissioned to carry out the FRC artificial intelligence research. They surveyed 103 organisations and interviewed 39 senior reporting and Investor Relations professionals.

The findings are a useful reality check, speaking to measured adoption, clear benefits in specific areas, and some governance gaps that boards should be aware of.

AI in Reporting: More Widespread Than You Might Think, But Adoption is Uneven

The research shows that 39% of organisations are already using generative AI (GenAI) in their corporate reporting, 31% are piloting it and 18% are considering adopting it within the year.

When we look a little closer, though, we can see that this adoption is uneven across industries and segments. Financial services businesses and larger companies are further ahead; smaller listed companies are more focused on basic efficiency tasks than widespread AI adoption for reporting.

One finding that may be most noteworthy for board members is that there’s evidence of a disconnect between actual use within teams and management or board awareness of that use. 

It’s clear that junior staff are using GenAI to create early drafts of reporting in a way that senior leaders don’t always know about. This may create risks that senior management aren’t alive to.

Where AI Is Actually Being Used

As with AI’s use elsewhere, AI in corporate reporting shows the clearest wins in process-intensive, lower-risk tasks: data extraction and reconciliation, anomaly detection, compliance checklists, cross-referencing disclosures against FRC guidance, and benchmarking the annual report against sector peers.

Investor Relations teams are finding GenAI genuinely useful for analysing earnings call transcripts, investor questions, and competitor announcements: distilling large volumes of unstructured information into themes and insights.

When it comes to crafting the narrative that sits around the data, those producing reporting are more cautious. While 61% of survey respondents said they use GenAI for drafting narrative, the research interviews told a different story — where it is used, it’s typically for first drafts and content like chart descriptions, not strategic commentary or forward-looking statements.

As in many other areas of work, AI use can be split into two areas: GenAI as co-worker, where it supports human-led workflow, and GenAI as co-creator, where it generates output independently. From the research, most reporting teams are currently firmly in the former camp.

The Risks of GenAI in Corporate Reporting

The research identifies three key risks when using GenAI for corporate reporting:

1. Accuracy. GenAI can hallucinate or misinterpret, which clearly has potentially serious consequences in a regulated, public-facing document.
2. Data quality. Aside from the accuracy issue, AI outputs are only as good as the underlying data. In many organisations, this remains fragmented, causing challenges in bringing the full picture to bear.
3. Transparency. If you pull reports together manually, you’re aware of the full process that created your data and conclusions. Some AI systems can’t easily explain their outputs, which is a problem when auditors or regulators ask questions.

These are very concrete and overt risks; there is also a subtler one, which doesn’t necessarily speak to inaccuracy or errors, but is important to the report’s audience (and therefore to your credibility). Maintaining a consistent, authentic management tone of voice is important; your investors and analysts will notice when the narrative sounds generic or different to the company’s actual story or usual style.

The research shows that only 44% of companies have mandated human oversight of AI-generated content, despite virtually everyone interviewed describing it as essential. This human oversight can be the difference between these risks being possible, and coming to fruition.

There is also a data governance risk that shouldn’t be ignored. In the research, 24% of GenAI users report using public AI tools alongside their own company platforms, potentially sharing corporate-confidential data in public AI platforms.

In the same way that using email to share your corporate information creates a risk, using open AI tools risks your information being vulnerable to bad actors.

What Good Governance of AI in Reporting Looks Like

Reporting on its research, the FRC is clear that it sees effective governance as vital. More than that, governance is the critical enabler of responsible AI adoption — rather than a brake on it.

With its ability to create early drafts and to gather, aggregate and analyse data, there’s no doubt that AI can play a role in reducing the stress inherent in the annual reporting process. But it does need some guardrails.

Companies taking the lead in using AI in this capacity have some common controls:

• Clear acceptable use policies
• Defined accountability for outputs
• Robust human review and validation at every stage
• Engagement with internal audit on AI use

For boards and audit committees specifically, this is increasingly an area of oversight responsibility — understanding how AI is being used in the reporting process, by whom, and with what controls. This is no longer optional but an essential element of corporate governance.

The FRC has indicated that it will publish case studies on AI use in reporting in the coming months, and we will share details of those as they become available.

AI + Human = Successful Corporate Reporting

It’s clear; AI is genuinely useful in corporate reporting, but the FRC’s research confirms that the areas where it adds most value are efficiency and process, not the narrative and judgment that define a great annual report.

The human element — experienced editorial oversight, regulatory knowledge, authentic tone of voice — remains what separates reports that simply comply from those that genuinely communicate.

This human element comprises both the experts within your business — the people who know your company and your data — and the external expertise you can tap into. Both are vital in enabling you to build on AI’s undoubted ability to support corporate reporting, providing the oversight and guardrails needed to make the most of GenAI’s potential.  

Perivan’s Shareholder Communications services bring that experience to the annual reporting process, supporting companies in creating accurate, engaging and user-focused corporate reporting.

Contact us to find out more.