Fraud Prevention # What if the scam came from inside a prison? Prison geointelligence: how the location of a digital transaction can reveal if it originated from inside a penitentiary facility — in real time, via API. +1.500 prisons mapped 4.757 cellphones seized in prisons (Op. Mute) R$ 4,9 bi lost to PIX scams in Brazil The problem ## Prisons have become operation centers for digital fraud The illegal use of smartphones inside Brazilian prisons has transformed penitentiary facilities into true **crime call centers**. From inside the cells, criminal organizations run PIX scams, fake kidnappings, social engineering, and digital fraud — all with access to the internet and banking apps. 2.2 million cases of virtual fraud in 2024 — 408% growth since 2018 Source: Brazilian Yearbook of Public Safety 2025 17.785 cellphones seized in Rio prisons alone between 2023 and 2025 Source: SEAP/RJ via O Globo 674 mil people in physical cells in Brazil — 3rd largest prison population in the world Source: SENAPPEN / SISDEPEN 2024 The scenario ## Two identical transactions. Completely different risks. Without prison geointelligence, your fraud engine can't distinguish a legitimate transaction from one originated inside a prison. Both arrive as a regular PIX. Transaction A ### Residential neighborhood — Joinville/SC **Operation:** R$ 3,500 PIX **Captured GPS:** -26.3044, -48.8487 **Nearest prison:** 14.2 km **Inside prison:** NO **% fraud inmates:** N/A (distant) → Transaction follows normal flow Transaction B ### Penal Unit "Ricardo Brandão" — Ponta Porã/MS **Operation:** R$ 3,500 PIX **Captured GPS:** -22.5302, -55.7162 **Nearest prison:** 0 m (inside) **Inside prison:** YES ⚠ **% fraud inmates:** 0.15% → Block, review, or escalate for analysis Without PrismaData, both transactions look identical to the fraud engine. With it, Transaction B is immediately flagged. The solution ## One API call. Complete prison context. Send latitude and longitude. Receive back the nearest prison, the distance in meters, whether the point is inside the geofence, facility type, prison population, and percentage of fraud inmates. GET /v1/location/prison?lat=-22.5302&lng=-55.7162 { "ponto_dentro_presidio": true, "distancia_borda_m": 0.00, "nome_estabelecimento": "Unidade Penal Ricardo Brandão", "tipo_estabelecimento": "PRESOS_PROVISORIOS", "municipio": "Ponta Porã", "uf": "MS", "pop_masculino_total": 658, "percentual_estelionato_total": 0.0015, "indice_lotacao_total": 0.49, "ambito": "ESTADUAL", "area_m2": 13732.88 } Key field ponto_dentro_presidio Direct boolean: was the operation done from inside? Distance distancia_borda_m Meters to the prison geofence boundary Estimated risk percentual_estelionato % of prison population incarcerated for fraud Applications ## At which customer journey moments to apply? Every digital operation that captures the device GPS is an opportunity to apply prison geointelligence as an additional risk signal. #### Onboarding Account opening from inside a prison? Strong signal for blocking or manual review. #### Login Banking app access from prison coordinates may indicate a stolen device or shell account. #### PIX / Wire Transfer Transfer originated from inside a prison: immediate blocking or hold for analysis. #### Account changes Email, phone, or address change made from inside a prison is highly suspicious. #### Credit request Loan request or credit limit increase from inside a prison: strong indicator of fraud. #### Password recovery / Device change One of the most exploited vectors by fraudsters. Prison location is a decisive signal. Integration ## How it works Simple integration via REST API. Just send the GPS coordinate captured at the time of the operation. Digital operation Captured GPS from device API PrismaData /v1/location/prison lat + lng Enriched response Inside? Distance? Which prison? Risk? Decision engine Score, policy rule or model feature REST Standard API < 200ms Typical latency Lote Batch endpoint available SDK Python available Coverage ## 1,500+ prisons with proprietary geofences We manually mapped the geographic polygons of over 1,500 prison facilities across Brazil — covering nearly all closed-regime establishments in the national penitentiary system. 27 Federal Units covered Geofences Manually created polygons, not centroids Exclusive Proprietary data — not available from any bureau Reference data: SENAPPEN / SISDEPEN records 1,424 official prison facilities (2023 data). PrismaData covers over 1,500 geofences, including mixed facilities and barracks variations within the same complex. Governance ## 100% public data. LGPD-compliant. Auditable. #### Public data All information is derived from official public sources (SENAPPEN, SISDEPEN, DEPEN). No personal data is processed. #### LGPD-compliant The API receives only geographic coordinates. It does not identify or track individuals. The enrichment is of the context, not the person. #### Auditable Every response is traceable and explainable. Ideal for supporting human decisions and regulatory compliance. ## Ready to add prison geointelligence to your fraud engine? Talk to PrismaData. We can run a pilot with your own coordinates in minutes. [ Talk to PrismaData ]() [ Explore the API → ]()