Nearly half of tax administrations surveyed by the OECD use artificial intelligence, while more than 80% are developing interfaces to exchange data with business software. The study of 108 economies shows tax digitalisation moving beyond online returns towards direct transfers of company data to the authorities.
By Adam Ujazdowski. Polish original published on 12 November 2026 at 11:58:38 UTC.
The OECD’s report, Digitalisation and Digital Transformation in Tax Administrations: Regional Trends and Insights, was published on 12 November 2026. It covers technologies already in use as well as those still being developed, with roughly three-quarters of administrations reporting a digital transformation strategy.
Fraud detection and risk assessment lead AI use
Among administrations using AI, 74% deploy it to detect tax fraud and evasion, and 70% use it for risk assessment. Algorithms can flag unusual tax filings and cases that warrant closer examination by officials.
Virtual assistants are used by 57% of AI-adopting administrations, while 49% use the technology to support administrative decisions. These percentages refer only to institutions using AI, rather than all 108 survey participants. The OECD does not say that AI is replacing officials or independently making final decisions.
More than 80% are developing APIs
More than 80% of surveyed administrations are developing application programming interfaces, or APIs, to connect government systems with business software. These interfaces can transmit tax returns, accounting data and payment information, although development does not mean that every service has been fully deployed.
Among administrations developing APIs, almost 60% provide external users with a catalogue of interfaces, and 65% work with third parties to design solutions. Such connections can reduce the need to re-enter information manually.
Digitalisation of tax administrations: OECD findings
Share of administrations reporting each practice
Source: OECD, 2026. Approximate figures; bar lengths are indicative. Developing APIs does not mean that they have been fully deployed.
Business systems feed data directly to tax authorities
Just over 60% of administrations report receiving data directly from business systems, including information beyond payroll. Of that group, 82% allow at least some data to be transferred automatically without human intervention.
Three-quarters of administrations also obtain information from third parties, such as financial institutions. More than 80% exchange large datasets with other public bodies, although the scope of these arrangements varies across economies.
Pre-filled returns extend beyond personal income tax
Tax administrations use information they already hold to prepare returns that taxpayers can check and approve. This is most widespread in personal income tax, particularly in some European countries.
Systems for value-added tax (VAT) and business tax filings are also being developed. Their operation depends on data from electronic invoices, accounting systems and other registers. This does not mean businesses in every jurisdiction already receive complete, automatically prepared returns.
AI adoption brings oversight requirements
About 75% of administrations have a digital transformation strategy, but individual projects are at different stages. Among institutions using AI, 85% report restrictions on its use, while 40% have an ethical framework for the technology.
The OECD reports that 81% of AI-using administrations check input data, 81% test system responses and 79% monitor results. These are self-reported practices, rather than independently verified measures of effectiveness.
The findings rely on administrations’ own reports
The 112-page report draws on the Inventory of Tax Technology Initiatives and primarily presents information for 2025 supplied by tax administrations themselves. The OECD says it did not independently verify those submissions.
The indicators cover different groups of respondents depending on the question. They therefore cannot be treated as comparable measures of technological effectiveness or used to infer increases in tax revenue or savings for administrations.
Source: OECD, Digitalisation and Digital Transformation in Tax Administrations: Regional Trends and Insights, 12 November 2026.
