evidoria

← Back to browse

Good practice

VeRa — Italy's AI risk-scoring algorithm for tax-evasion detection

Italy · Rome · See the Italy profile

Italy's Agenzia delle Entrate deploys VeRa, an AI algorithm cross-referencing declarations with bank data, property records and electronic payments. In 2022 it flagged >1 million high-risk filings, preventing >€6.8 billion in VAT fraud; legal basis in Law 111/2023.

1+ million
High-risk VAT cases flagged (2022)
6.8+ € billion
Fraud prevented (2022)
20.2 € billion
Total recovered from tax-evasion enforcement (2022)
VeRa — Italy's AI risk-scoring algorithm for tax-evasion detection

Details

Maturity
Established
Promoter
Agenzia delle Entrate (Italian Revenue Agency)
Period
2022-present
Keywords
tax compliance, machine learning, risk scoring, VAT fraud detection, e-invoicing, big data

Context

VeRa (Verifiche e Riscontri Automatizzati) is Italy's AI tax-risk-scoring algorithm, run by the Agenzia delle Entrate with state ICT company Sogei, reaching operational scale in 2022. It ingests e-invoices, bank-account data, the property register, motor-vehicle records and customs filings to compute a composite risk score per taxpayer.

Activities

Risk scores are used to prioritise audits and voluntary-disclosure nudges; personally identifying data are pseudonymised before analysis; an AI chatbot was added in 2023 to flag anomalies in real-time VAT returns for human officers. Law no. 111/2023 gave the Agency formal statutory authority to use AI for compliance detection.

Results

In 2022 VeRa flagged more than one million high-risk VAT cases; the Italian Revenue Agency attributed prevention of more than €6.8 billion in fraud that year, against €20.2 billion recovered overall from tax-evasion enforcement.

Conclusions

The European Commission's Technical Support Instrument provided independent documentation of the approach; similar risk-scoring concepts exist in other OECD countries, though VeRa itself has not been formally transferred to another jurisdiction.

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful