Transparency in Colombian public procurement has an unexpected enemy: the state's own institutional architecture. This is not only about corrupt officials or weak political will, but about a system in which the mechanisms of opacity are built — sometimes inadvertently — into the very design of the entities that handle the largest sums.
The case of the National Disaster Risk Management Unit (UNGRD) is instructive. Since at least January 2026, civil society organisations and news outlets have documented the entity's refusal to publish complete information on its highest-value contracts, despite a court order requiring it. The institutional response — invoking exceptions, stalling through appeals and complying only in part — repeats a pattern documented at other entities with similar profiles: wide discretion, large budgets and little day-to-day media scrutiny.
Three risk vectors converge in this picture.
The first is the selective opacity of entities such as the UNGRD. When an entity breaches transparency mandates and faces few consequences, the message sent to the system is that accountability is optional. SECOP II data show that the UNGRD has awarded contracts worth trillions of pesos using the declared-emergency route — a legitimate tool for emergencies — but whose repeated, systematic use is a statistical indicator of possible abuse. LuxIA detects this pattern by comparing an entity's frequency of declared emergencies with the sector average and with its own institutional track record.
The second vector is the migration of the SECOP system. The transition to a new state procurement platform, valued at roughly COP 23,497 million, creates what digital-government specialists call windows of structural opacity: periods in which historical data become hard to compare, contracts may be misclassified and the audit trail breaks. Such periods have historically been used to award contracts with less public visibility. Monitoring time series — before, during and after the migration — is essential to catch anomalies that would otherwise stay invisible.
The third vector, and perhaps the most pressing, is electoral. With the 2026 presidential elections on the horizon, off-budget funds take on particular importance. Unlike contracts that pass through the national budget, these funds operate with less legislative scrutiny and can concentrate significant resources in specific regions under investment formats that, on the face of it, have nothing to do with electoral politics. Transparencia por Colombia has documented how these mechanisms have historically been used to build territorial networks of influence, especially in the 18 months before an election.
What connects the three vectors is the gap between the availability of data and the capacity to interpret it. Colombia has one of the most developed open procurement data systems in Latin America — SECOP II is a powerful tool — but the sheer volume of information available far outstrips the analytical capacity of oversight bodies and civil society. In 2025, more than 1.2 million procurement processes were published on the system. No human team can review them all.
This is where artificial intelligence applied to public data stops being a technological luxury and becomes a democratic necessity. LuxIA cross-checks SECOP II data in real time against early-warning indicators: concentration of contracts in a single supplier, runs of direct awards without documented justification, contracts signed in the 72 hours before the close of the fiscal year, and abrupt shifts in an entity's contracting profile compared with earlier periods.
The UNGRD case is not exceptional: it is representative. For every entity that makes the headlines, there are dozens whose risk patterns go undetected — not because the data do not exist, but because nobody has had the time or the tools to read them. Changing that is the reason LuxIA exists.