Good practice

CIAT e-IAD — Detetor de anomalias IA para fraude em faturação eletrônica na Costa Rica

Costa Rica · San José · See the Costa Rica profile

A autoridade fiscal da Costa Rica usa o modelo ML open-source CIAT e-IAD para detetar anomalias em faturas eletrônicas. Em 2026, o sistema expoz uma rede de fraude de ~₡80 mil milhões com ~200 contribuintes e recuperou ₡8 mil milhões em 2025. O CIAT replicou-o na Guatemala e no Rio Grande do Sul.

~200

Contribuintes implicados na rede de fraude desmantelada (by April 2026)

~₡80 billion (~USD 150 million)

Valor dos comprovativos falsos identificados (by April 2026)

~₡8 billion (~USD 15 million)

Receita fiscal recuperada (2025)

Inspetores da DGT responsáveis pelos maiores contribuintes

12 months

Período de desenvolvimento do modelo

CIAT e-IAD — Detetor de anomalias IA para fraude em faturação eletrônica na Costa Rica

Details

Maturity: Established
Promoter: Dirección General de Tributación, Ministerio de Hacienda
Period: 2022–present
Keywords: tax compliance, e-invoicing, unsupervised machine learning, anomaly detection, CIAT

Context

A Dirección General de Tributación da Costa Rica, em conjunto com o CIAT e a Microsoft, implementou o e-IAD, um modelo de aprendizagem automática não supervisionada e de código aberto, co-desenvolvido ao longo de 12 meses por especialistas fiscais de sete países membros do CIAT, que analisa milhões de faturas eletrónicas, declarações fiscais e dados do registo de contribuintes para sinalizar anomalias e padrões de comportamento compatíveis com evasão fiscal organizada.

Results

Até abril de 2026, o sistema ajudou os investigadores da DGT a desmantelar uma rede de fraude com cerca de 200 contribuintes, envolvendo comprovativos falsos num total aproximado de ₡80 mil milhões (~USD 150 milhões); só em 2025, auditorias assistidas por IA recuperaram cerca de ₡8 mil milhões (~USD 15 milhões) deduzidos de forma irregular através de faturas eletrónicas não fundamentadas. Desde então, o CIAT implementou a mesma solução na Guatemala e no Rio Grande do Sul, Brasil.

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.

C-NAPSE web research — accessed 2026-06-21
Dirección General de Tributación, Ministerio de Hacienda — accessed 2026-06-21
Microsoft customer story — CIAT e-IAD in Costa Rica (June 2023) — accessed 2026-06-21
El Financiero CR — AI exposes ₡80 bn fraud scheme (April 2026) — accessed 2026-06-21

Attachments

Source image

Quality score: 73/100

How is this calculated?

Evidence: Observational / pre–post

Top 16% in this catalogue (246 scored practices)

Scores cluster high, so position within the catalogue is often more telling than the number alone.

Evaluation

Evaluated by C-NAPSE

Score profile — the dimensions behind the score — how it's computed

Evidence of impact / measured public value2/3

₡80 billion fraud scheme (~200 taxpayers) and ₡8 billion recovered in 2025 are documented by El Financiero CR (April 2026) and confirmed by CIAT’s official e-IAD project page.

Transparency, fairness & accountability2/3

The e-IAD is open-source and co-developed by seven CIAT countries with a published methodology; individual audit decision rationales are not yet fully disclosed to taxpayers.

Transferability / demonstrated replication3/3

CIAT has already replicated the e-IAD in Guatemala and Rio Grande do Sul (Brazil) in addition to Costa Rica, confirming replication across three different jurisdictions.

Scalability beyond pilot2/3

Deployed at national scale processing millions of electronic invoices; operational effectiveness remains constrained by DGT’s small inspector workforce (30 inspectors for major taxpayers).

Governance, capability & sustainability2/3

Director General of Taxation is publicly accountable; CIAT provides multilateral oversight; the open-source model enables external scrutiny, but no formal taxpayer redress mechanism was documented.

Cite this

Evidoria. CIAT e-IAD — Detetor de anomalias IA para fraude em faturação eletrônica na Costa Rica. Persistent ID: b8d58328-36cc-4a48-83cd-605503ab7bca.

Export as decision brief (PDF)

Similar practices you may find useful

★ 80

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

Italy

Italy's Agenzia delle Entrate deploys VeRa, an AI algorithm cross-referencing declarations with bank data, property records and electronic payments. In …

★ 87

Revenue REAP — Ireland's AI risk engine for human-supervised tax-compliance targeting

Ireland

Since 2011, Ireland's Revenue Commissioners have used AI risk scoring (REAP) to target tax-compliance interventions; 290,000 data-analytics-driven interventions in 2023 …

★ 80

ESCORT — Latvia's State Revenue Service ML risk-scoring system for tax compliance

Latvia

Latvia's SRS uses ESCORT ML to score every registered company for fraud risk across tax and labour domains. Fraud detection …

★ 80

SARS AI compliance engine — South Africa's machine-learning tax risk platform

South Africa

South Africa's SARS uses ML models across 1.2 billion annual records to select audits and detect refund fraud. In FY2024/25 …