NAV Online Számla — Hungary's real-time invoice reporting and AI-assisted VAT fraud detection
Hungary
Hungary requires all businesses to report invoices to NAV in real-time XML since 2021; ML risk-scoring targets suspicious companies before …
Malta · Valletta · See the Malta profile · See the Valletta profile
Evidence: Observational / pre–post Top 14% 80/100 · Ask Evidence Copilot about this practice
Malta committed €3m to deploy SAS AI tools that cross-reference registries and bank data to flag VAT fraud; officials attribute most of a €650m year-on-year 2024 revenue rise to the tools, with VAT refunds now processed within a month instead of three to four.
MT001Taxes and social contributions made up 86.7% of Malta's 2022 government revenue (€5.1 billion). The Malta Tax and Customs Administration (MTCA) previously relied on manual sampling to select audit cases, leaving most filings unchecked, prompting Finance Minister Clyde Caruana to commit €3 million to AI software already used by tax authorities in the UK, New Zealand, the Netherlands, Ireland and Canada.
Automatically flag missing-trader ('carousel') fraud, fictitious invoices and inflated VAT refund claims by cross-referencing taxpayer registries, bank-account data and property, land and vehicle records, replacing manual sampling.
MTCA rolled out SAS Visual Investigator, SAS Visual Analytics, SAS Intelligent Decisioning and SAS Viya 4, initially scoped to VAT returns and drawing on international data-sharing agreements, with extension to all tax forms planned within three years.
Malta reported collecting €650 million more in tax in 2024 than in 2023, with officials attributing most of the increase directly to the AI tools. VAT refunds that previously took three to four months are now processed within the same month.
These figures are self-reported by the tax administration and its vendor rather than independently audited, and the claim that 'most' of the 2024 revenue increase stems from AI has not been isolated from other concurrent enforcement measures or economic conditions.
National / regional programmes
Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.
Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.
Where this practice's information was retrieved from, and when.
Hungary
Hungary requires all businesses to report invoices to NAV in real-time XML since 2021; ML risk-scoring targets suspicious companies before …
Slovenia
Slovenia's FURS runs two ML risk-scoring models in its eDIS platform — one flags irregular VAT-refund claims since 2019, another …
Rwanda
A study built with Rwanda Revenue Authority data used MiniLM, UMAP and HDBSCAN to harmonize 4.1 million multilingual product names …
United States of America
The IRS grew from 10 to 126 active AI use cases between 2022 and mid-2025 to flag high-risk tax returns …
Open full copilot Grounded in cited practices — always check the sources.