evidoria

← Back to browse

Good practice Imported

DNCP's AI-Assisted Anomaly Detection — Paraguay's Procurement Watchdog Flags Bidding Risks

Paraguay · Asunción · See the Paraguay profile · See the Asunción profile

Evidence: Descriptive / self-reported Top 83% 40/100 · Ask Evidence Copilot about this practice

Paraguay's procurement authority DNCP has run a government AI system since 2022 to flag bidding anomalies and predict corruption risk, reconfirmed active in 2025 by the Vice Minister of Technologies — though no detection-accuracy figures are public.

284 million PYG (~US$40,000)
AI-layer contract value (June 2022)
100+ statistics
Dynamic statistics published on procurement portal (2025)
DNCP's AI-Assisted Anomaly Detection — Paraguay's Procurement Watchdog Flags Bidding Risks

Details

Maturity
Established
Promoter
Dirección Nacional de Contrataciones Públicas (DNCP)
Period
2022–2025
Keywords
public procurement, anti-corruption, digital government

Context

Paraguay's Dirección Nacional de Contrataciones Públicas (DNCP) runs the country's single public-procurement portal, already publishing over 100 dynamic statistics and frequently cited among the world's most transparent procurement systems.

Activities

DNCP awarded a multi-year contract in June 2022 (₲284 million, roughly US$40,000) titled 'Mejora continua de la aplicación de inteligencia artificial al proceso de contrataciones públicas' to build and maintain an AI layer over its tender data, used to detect anomalies in tenders, predict risks, and improve citizen service through an in-development virtual assistant.

Results

In November 2025, Vice Minister of Technologies Klaus Pistilli confirmed the system remains active, as part of a broader national plan to expand AI use across government.

Conclusions

The approach mirrors AI-based procurement-oversight systems already running in Ukraine, Colombia and Bulgaria, but Paraguay's deployment remains thinly documented — no detection rates, false-positive rates, or independent evaluation have been published, and an earlier civic open-data bot (@controlcivicopy, 2021) suggests anomaly-flagging on the same dataset is technically straightforward, a benchmark the official system's public track record has not yet matched.

Implementation
Indicative cost
Low (< €50k) — ₲284 million (~US$40,000) multi-year contract awarded June 2022, described as ongoing/active through November 2025.
Time to results
Medium (1–3 years) — Contract awarded June 2022, reconfirmed active November 2025 (~3.5 years of continuous operation reported).
Staffing & skills
Dirección Nacional de Contrataciones Públicas (DNCP), Ministry of Technologies (Vice Minister Klaus Pistilli, public spokesperson)

Conditions for success

  • An already highly transparent, single national procurement portal to build the AI layer on top of
  • A sustained multi-year contract rather than a one-off pilot
  • Alignment with comparable procurement-oversight AI systems in Ukraine, Colombia and Bulgaria

Common failure modes

  • No detection rate, false-positive rate, or independent evaluation has been published
  • The AI model's flagging criteria are not disclosed despite DNCP's otherwise strong data transparency
  • A prior civic open-data bot built by two individuals suggests the underlying anomaly-flagging task is not technically demanding, raising questions about the marginal value delivered by the funded system

Where it fits

Governance type
national government agency
Scale
national (single procurement portal)
Income level
upper-middle-income

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

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

Attachments

Similar practices you may find useful