From Filed Accounts to Machine-Readable Finance: The UK Reporting Infrastructure Shift
Merixa Insights · Transformation · IFRS Reporting
The UK is changing not only how company accounts are submitted, but how financial information can be read, compared and analysed. For finance teams, this is an infrastructure change before it is a filing change.
From April 2028, all UK-registered companies will be required to file annual accounts through commercial software in iXBRL format. Companies House will close its web and paper routes for accounts filings. Small companies and micro-entities will file profit-and-loss accounts, although they may opt out of public publication. Abridged accounts will be removed, audit-exemption statements strengthened and the component parts of filed accounts submitted together.
iXBRL combines a human-readable report with machine-readable tags. Companies House states that the reform will improve financial-data quality and create greater opportunities for accounts information to be aggregated, compared and analysed. The Financial Reporting Council’s draft 2027 Taxonomy Suite shows the technical layer developing alongside the reform, with updates for IFRS 18, amendments to IFRS 9 and IFRS 7, FRS 102, FRS 103 and relevant auditing standards.
The transformation begins before the final accounts file is produced. A tagged report is only as reliable as the accounting structure, classifications and controlled data flow beneath it.
Workstream one — data architecture
The chart of accounts, consolidation mapping, statutory disclosures and management-reporting dimensions should be aligned sufficiently to reduce repeated manual reclassification at year end.
Where the management ledger and statutory reporting structure operate through different definitions, tagging may reproduce the inconsistency rather than resolve it.
Workstream two — taxonomy readiness
Finance teams need to establish which disclosures map directly to standard taxonomy elements, where extensions may be required and how tagging judgements will be reviewed.
Tagging should not become a separate technical exercise detached from accounting ownership. The judgement remains connected to the meaning, classification and presentation of the underlying financial information.
Workstream three — evidence and lineage
The organisation should be able to trace a reported amount from the filed statement through consolidation entries and adjustments to the originating ledger and supporting evidence.
This becomes more important when machine-readable information can be compared across periods, companies and reporting frameworks at scale.
Workstream four — control redesign
Preparation, tagging, review, software validation, filing approval and post-submission correction require named ownership and retained evidence. Existing year-end checklists may not cover these responsibilities adequately.
The wider AI environment raises the standard further. HM Treasury’s Financial Services AI Adoption Plan proposes an industry repository for AI incidents and near misses, assessment of important AI and cloud providers under the Critical Third Parties regime, and a voluntary assurance framework supporting independent assessment and certification of third-party AI systems.
techUK separately reported that 61% of surveyed firms were using or assessing generative AI and 42% were using or assessing agentic AI in 2025. Its report included production deployments in areas including fraud detection, customer operations, capital-markets research and insurance risk assessment.
These developments connect machine-readable reporting with machine-assisted analysis and increasingly autonomous financial processes. The practical transformation is not to add AI to the existing close. It is to establish a controlled architecture in which data definitions, interfaces, reporting logic, evidence and technology assurance operate as one environment.
The implementation questions
- Can the ledger support consistent statutory and management classifications?
- Are consolidation adjustments controlled and traceable?
- Who owns taxonomy judgements and software validation?
- Are AI and cloud dependencies included in third-party risk assessment?
- Can the organisation reproduce the evidence behind each material filed output?
April 2028 is the filing date. The operating-model work belongs earlier.
Merixa supports organisations in redesigning finance processes, reporting architecture, data flows and control ownership before digital reporting requirements expose structural weaknesses.