Insights
TAX & BUSINESS
Artificial Intelligence in the Tax Administration
The use of artificial intelligence (AI) helps the tax office to generate additional tax revenue. In 2024 the authority was able to secure around EUR 354 million in extra tax receipts through the use of state‑of‑the‑art analytical methods.
The Predictive Analytics Competence Center (PACC) is responsible for this, a specialised unit within the Ministry that systematically detects tax fraud and compliance breaches using machine learning and data analysis. In the past year approximately 6.6 million cases were examined by the risk models. These include income‑tax, corporation‑tax and VAT cases, as well as employee tax assessments. In addition to obvious tax fraud, roughly 23.4 million cases were reviewed for possible compliance violations, in particular concerning abusively claimed subsidies and benefits and regarding shell companies.
AI as a key to tax fairness
The specialised unit employs, among other things, text mining, machine learning and AI‑driven risk models to identify anomalies early. It draws on a wide range of data sources – from historic tax data to external databases.
An example of the working method is provided by a recent case: During a tax audit it was discovered that a farm had purchased machinery but had not recorded it in the asset register. The high‑value agricultural vehicles were sold after years of use without the sale proceeds being accounted for. The loss amounted to EUR 300,000 in unpaid duties. The detection was made possible by a targeted data match between the vehicle registration database and the farm’s financial data.
Increasing efficiency
According to a press release from the Ministry of Finance, advances in automated case selection not only improve hit rates but also increase the efficiency of the entire tax administration. Especially in the areas of tax audits, Joint Inspection of Wage Contributions and Levies (GPLB) and customs, AI models are used to identify, with a high probability, costly but lucrative examinations. In addition to traditional tax audits, the PACC also deals with future topics such as cryptocurrencies, international data exchange and the standardisation of audit procedures.
Co‑operation with national and international partners
For the coming years the Ministry of Finance plans to further expand the use of AI, in particular by integrating generative AI such as large language models (LLMs). Moreover, cooperation with national and international partners is to be deepened in order to recognise risks even faster, possibly in real time. The aim is to further increase the effectiveness and impact of the tax administration.
AI as a key to tax fairness
The specialised unit employs, among other things, text mining, machine learning and AI‑driven risk models to identify anomalies early. It draws on a wide range of data sources – from historic tax data to external databases.
An example of the working method is provided by a recent case: During a tax audit it was discovered that a farm had purchased machinery but had not recorded it in the asset register. The high‑value agricultural vehicles were sold after years of use without the sale proceeds being accounted for. The loss amounted to EUR 300,000 in unpaid duties. The detection was made possible by a targeted data match between the vehicle registration database and the farm’s financial data.
Increasing efficiency
According to a press release from the Ministry of Finance, advances in automated case selection not only improve hit rates but also increase the efficiency of the entire tax administration. Especially in the areas of tax audits, Joint Inspection of Wage Contributions and Levies (GPLB) and customs, AI models are used to identify, with a high probability, costly but lucrative examinations. In addition to traditional tax audits, the PACC also deals with future topics such as cryptocurrencies, international data exchange and the standardisation of audit procedures.
Co‑operation with national and international partners
For the coming years the Ministry of Finance plans to further expand the use of AI, in particular by integrating generative AI such as large language models (LLMs). Moreover, cooperation with national and international partners is to be deepened in order to recognise risks even faster, possibly in real time. The aim is to further increase the effectiveness and impact of the tax administration.