Optimising reporting for AI / Section 6:
How digital-first improves quality
A digital-first approach involves taking the best aspects of a well-established control process, enhancing them and then giving the management team more time to apply those controls.
What is the difference between digital-first and print-first formats?
In whatever format they choose, readers of a corporate report need the content to be high quality to trust what they read.
A report in digital format does not, of itself, guarantee a high-quality report. Certainly, if all of the information is in HTML format and available for AI to find, a reader’s access and experience is much improved. But there’s more to it than this.
High-quality digital-first reporting must also retain the best qualities of the print and PDF formats that have evolved over decades to make reports easy to read and understand. These qualities include consistent structure, clear narrative flow and easy to navigate design.
And robust quality controls now matter even more than ever. Digital reporting content errors and format readability problems can be discovered in seconds by machines. This prompts companies to add digital quality controls to their processes.
60x
A rise in user visits of 60x compared to the equivalent PDF-only, pre-digital reporting.
10x
Rise in the total number of PDF downloads.
Different aspects of quality
Consider how an annual report’s content structure, which follows consistent conventions, translates to digital. A typical table of contents in a print or PDF report has four primary sections: Strategic report, Governance, Financials, Other information (or a similar variant of these). In digital-first, this structure also translates perfectly into an online navigation structure.
Using either navigation or search and AI tools, a reader can journey to the precise content they need, without the need to scroll or manually search through a lengthy PDF document.
Digital also offers many opportunities for reporting content to be enhanced, and digital analytics show that this significantly improves user engagement. Two years of user data from one early adopter of digital first reporting showed a rise in user visits of 60x compared to the equivalent PDF-only, pre-digital reporting. This is the result of the reports’ increased visibility in both search engines and AI tools. Interestingly, this also delivered a 10x rise in the total number of PDF downloads too, demonstrating how online and PDF reports are complementary and meet different needs.
Videos, interactive charts and animation can all be included within a digital-first annual report, alongside its static equivalent in the PDF, print and regulatory filing versions. These enhanced features make content much easier to access, digest and understand, without any change to the underlying disclosures.
Annual Reports carry a very broad range of information. When people ask AI any question about a company, whatever that question, a full online report gives AI a far better chance of finding and serving up accurate information – sourced directly from the company, rather than from a third party website. And, as outlined in section 5, when the online report is enhanced with AI-friendly JSON-LD and iXBRL tagging, the accuracy of AI’s answers is transformed.
Digital automations like these solve data consistency risks and real-world pain points for investor relations, communications and reporting teams.
How digital processes improve reporting quality
Many digital-first design processes are the same or similar to existing print-first design processes. For example, the digital-first process still retains the PDF report that can be produced when the project reaches a milestone, which means that editing and proofing controls can remain relatively unchanged. Many management teams still consider PDFs to be the most effective way to check content and in a digital-first approach they can do so.
However, many of the manual and risky parts of the traditional design and production software process are being improved by digital software automations.
For example, page cross referencing and numbers – all cross references in digital versions become hyperlinks, and these hyperlinks are automatically assigned page numbers when the PDF and print versions are exported. This process also builds the table of contents automatically into an interactive menu in the PDF version. Page cross references and numbers update whenever a change is made so no manual editing of page numbers is required.
Other valuable automations include the use of recurring “key content” and “sync to source data”. These features connect data from finance systems, and can link content which appears multiple times in a report, so all instances of this common content are automatically updated.
For example “Profit for the year” might first appear early in an annual report on a highlights page, and then be referred to in reports made by the chair, CEO and finance director, then appear in the profit and loss account as part of the financial statements. With digital reporting software, all recurring figures or text content can be linked to a ‘parent’ item – so that each ‘child’ instance of the same data updates simultaneously to match the parent.
These automations can also link common content and data across a whole suite of reports and announcements, all from the single parent source – making the last mile of regulatory and investor communications more efficient.
Digital automations like these solve data consistency risks and real-world pain points for investor relations, communications and reporting teams.
You can see some of these automations in action in this three-minute video.
More time means better control... and better quality reporting
The digital-first reporting process offers a vital advantage for improving your controls – more time for the necessary checks and controls to be applied. This is because in a digital-first process, preparation and review of many elements can be started earlier in the process. It streamlines the volume of controls tasks required in the last mile sign-off period, reducing the risk of errors that arise when people are rushed, tired or both.
A good example of this is in ESEF reporting controls. In digital-first software like Reportl, XBRL tagging is fully integrated into the system (not in separate software). This means the structural XBRL tags can be applied much further in advance of the year-end, with most tagging decisions made at an earlier stage. Tagging is ‘structural’ so, while the text and data within each tag may be edited right up to sign-off, the tagging structures remain constant.
This solves the many well-known quality and timing challenges associated with the PDF to ESEF conversion process.
Regulators have highlighted several areas in reporting quality that require attention.
Less time for controls… means greater risk of errors
When a PDF is used as the source format for ESEF, even if the tagging is prepared with earlier PDF proofs, the tagging must be applied in full again after the final PDF is signed off. This increases the potential risks of unintended tagging changes, and these are difficult to identify between versions, so require slower, more manual checks.
Combined with tight filing deadlines, the PDF conversion method reduces time for key tagging decisions and compliance controls. In some situations, there is no time left for adequate tagging and format checks, which significantly increases the risks of data errors in the report.
Regulators have highlighted several areas in reporting quality that require attention. In April 2025 the FRC announced that they had begun reviewing the quality of ESEF digital reports, and provided useful guidance on improvements needed.
Financial and regulatory reporting controls are enhanced
Digitisation is exciting, but it’s far from being a leap of faith. It is not only enhancing user experience and AI readability, digitisation is also transforming the robustness of the data, traceability of sources and the process controls in report creation.
Regulators have been preparing for the rise in demand for higher quality machine-readable data for some years.
In the UK, a requirement known as Provision 29 has been introduced. Directors are required to make a declaration that material controls, including regulatory compliance controls are effective. Already we are seeing listed companies place their annual report controls, including those related to ESEF digital reporting, firmly within the scope of activities which support the declaration. This is prudent because, should a material misstatement be found in an annual report, hindsight can be applied to indicate that there must have been a material control failure that caused the material misstatement. Read an interpretation of what Provision 29 might mean for digital reporting controls.
Regulators have explicitly advised on the risks and limitations of the print-first PDF conversion process for ESEF. Anticipating the challenges that lay ahead for companies, in 2022 the FRC first advised that a native HTML (digital-first) process for ESEF is the best practice way to achieve responsive, interactive, accessible and SEO-optimised reporting, and they continue to maintain that advice. In May 2025, ahead of Provision 29, regulators repeated their advice that digital reports must be adequately checked and they highlighted the key areas for attention.
Unreliable data today increases the risk of AI errors and potentially mistaken investment decisions far into the future.
AI magnifies the potential risks of poor digital reporting data
Finally, it is worth noting that non-compliance is not the only risk from publishing poor data. The digital data in mandatory ESEF reporting will remain visible to AI tools for decades. This means unreliable data today increases the risk of AI errors and potentially mistaken investment decisions far into the future.
Explaining aspects of quality and controls in digital-first reporting is hopefully useful for those considering a change. In our final section we make some predictions about AI and digital-first reporting in the future.