evidoria

← Back to browse

Good practice Imported

Peru's IRI — an AI Risk Index Scoring 190,000+ Public Investment Projects for Corruption and Mismanagement Signals

Peru · Lima · See the Peru profile · See the Lima profile

Evidence: Observational / pre–post Top 36% 67/100 · Ask Evidence Copilot about this practice

Peru's Contraloría General launched the IRI in March 2024, an AI index scoring 190,000+ public investment projects nationwide against 24 indicators from 10 state databases, flagging ~40% of projects at three key ministries as high financial or governance risk.

190,000+
Public investment projects scored nationwide
~40 %
Share of projects flagged high-risk at three key agencies
2,200
Barred suppliers identified via AI (2023)
S/584 million
Value of contracts secured by barred suppliers (2023)

Details

Maturity
Scaling
Promoter
Contraloría General de la República del Perú
Period
Mar 2024 – ongoing
Keywords
public investment oversight, anti-corruption, audit, machine learning, public finance

Context

Peru's Contraloría General de la República (CGR), the country's supreme audit institution, has invested in AI, big data and machine learning for 'control concurrente' (real-time government oversight), building on a 2023 effort that used AI to detect 2,200 suppliers barred from state contracting who nonetheless secured public contracts worth S/584 million, plus roughly 700 disqualified individuals still working in public service.

Activities

On 20-21 March 2024 the CGR presented the Índice de Riesgos en la Inversión Pública (IRI), described as Peru's first platform to evaluate more than 190,000 viable and active public investment projects nationwide in real time, applying AI algorithms across 24 objective financial, budgetary, operational and control indicators drawn from roughly ten official state databases.

Results

The CGR's first results flagged the regions of Áncash, La Libertad, Junín and metropolitan Lima as holding the most high-risk projects, and found that roughly 40% of projects run by the Health Ministry, Agrorural and Provías Nacional carried high risk.

Conclusions

The IRI's own descriptions frame it as a diagnostic, risk-flagging layer rather than causal proof of corruption or waste in any single project, and its effect on actual investment execution or graft-recovery outcomes has not been independently measured since its March 2024 presentation.

Implementation
Indicative cost
Medium (€50k–€500k) — Built on an existing national audit institution's data-interoperability investment across ten state databases; no separate IRI-specific budget figure is published.
Time to results
Medium (1–3 years) — Presented in March 2024 and already operating in real time across the full national portfolio of 190,000+ projects -- fast to full-scale, though still a young tool without a multi-year track record.
Staffing & skills
Contraloría General de la República (CGR) data and audit analysts

Conditions for success

  • interoperable access across roughly ten official state databases to build the 24-indicator risk profile
  • a constitutionally mandated supreme audit institution with authority to publish findings and name high-risk agencies
  • a defined set of objective, replicable indicators covering financial, budgetary, operational and control risk

Common failure modes

  • the index is explicitly a risk-flagging diagnostic, not proof of corruption in any single project, so treating it as a verdict would be a misapplication
  • no independent evaluation yet shows the IRI has changed investment execution, delay or graft-recovery outcomes

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.

Similar practices you may find useful