SAT's Use of AI Slashes Collection Costs by 21.4% in 2023

Mexico’s Tax Administration Service (SAT) is leveraging artificial intelligence to identify taxpayers who evade taxes or engage in improper tax planning. This technology has led to a 21.4% reduction in the cost of tax collection. In 2023, the cost was MX$0.28 per MX$100 collected, down from MX$0.34 in 2019. 

All data generated from electronic invoices or digital tax receipts (CFDI) is stored in a database and analyzed using public servers in combination with Machine Learning and Graph Analytics algorithms. This approach enables the efficient handling of large data volumes and facilitates the detection of patterns and anomalies that may indicate improper fiscal behavior. From January to June 2024, SAT collected MX$2.5 billion (US$134.3 billion), surpassing the Federal Revenue Law (LIF) budget and representing a 6.1% increase compared to the same period in 2023.

Gari Flores Hernández González, general administrator of Tax Collection, SAT, explained that while AI is a powerful tool, final decisions are always made by public servants. “AI helps us manage large amounts of information, differentiating between compliant taxpayers and those who are not. This allows us to be more efficient and allocate our human and financial resources more effectively,” she said.

AI’s application at SAT has proven effective in identifying taxpayers involved in improper tax planning or issuing false invoices. For instance, Flores mentioned an example where AI detected an inconsistency, such as a funeral home in Mexico City issuing invoices for bicycles sales in Tijuana. These discrepancies are flagged for further review by human auditors.

The integration of these advanced technologies into SAT's processes is part of the “Master Plan” initiative announced in January this year. This strategy aims to modernize and optimize tax auditing and taxpayer service through advanced technologies. 

SAT has indicated that AI will be used to review areas such as vulnerable activities, improper applications of VAT credit balances, foreign trade taxes, internal taxes, the fuel market, simulated operations, pension simulations, and payroll outsourcing.

“This technology will target sectors such as automotive, alcoholic beverages and tobacco, construction, pharmaceuticals, hydrocarbons, logistics, technology platforms, real estate services, insurance and financial services, and transportation,” detailed SAT.