With the new Government in office, nothing should prevent the Colombia Compra Eficiente Agency from issuing guidance for public entities to begin adopting Artificial Intelligence in government procurement.
There is no general constitutional, legal, or regulatory prohibition against using AI in public procurement in Colombia, provided that legally relevant decisions remain with competent public officials and the principles governing public procurement are fully observed.
But a reality check is needed: not every public entity is ready to deploy AI immediately. Each entity must first assess whether it has the necessary technological capabilities—hardware, software, data architecture, connectivity, cybersecurity, interoperability, and human capacity—to use AI responsibly and effectively. This is a prerequisite for implementation, not a legal barrier to adoption.
AI deployment must also respect explainability. AI-assisted decisions, recommendations, or alerts must be understandable, traceable, and justifiable in human terms, allowing public officials to review the analysis, validate its conclusions, and assume full responsibility for any resulting decision.
If the legal framework allows it, and if technological prerequisites can be progressively addressed, Colombia should move forward as soon as practicable.
The reason is straightforward: public procurement is one of the most information-intensive functions of government and one of the areas where AI could generate significant practical benefits.
A single procurement process may involve hundreds—or thousands—of pages of legal, technical, and financial information: market studies, previous contracts, technical specifications, risk matrices, qualification documents, evaluation criteria, observations, and regulatory requirements.
Much of this work is repetitive, time-consuming, and highly structured. That is precisely where AI can add value.
The objective is not procurement without humans. It is better procurement because competent public officials are better equipped.
AI can help process information faster, detect inconsistencies and omissions, compare large volumes of documentation, identify risks, and improve the quality and traceability of procurement analysis.
The discussion must now move from general feasibility to concrete implementation: where can AI be responsibly deployed across the procurement lifecycle—from planning and structuring to evaluation, contract execution, monitoring, and closeout—while preserving human oversight, legal accountability, and the authority of competent public officials and government entities? That will be the focus of my next publication.
This evolution should be progressive, practical, and responsible—supporting public servants without displacing the professional judgment that cannot, and should not, be delegated to machines.