Random Monkey (Datasketch) has selected the Resident Coordinator Office (RCO) project in Colombia as its new pilot partner and use case. This new use case still allows the team to transfer the learnings, tools, and technical approach developed through its previous work on monitoring and evaluation processes with multiple UN agencies in Colombia.
Random Monkey is developing an AI-powered data platform for the Resident Coordinator Office in Colombia to consolidate fragmented UN project information, improve data governance, and support faster, evidence-based coordination across agencies
The project focuses on improving the RCO’s knowledge bases and data governance in Colombia. Random Monkey has already engaged with the RCO team and can align the work with UNICEF Colombia’s monitoring and evaluation team to ensure the solution responds to country-level coordination needs.
Updated investment thesis
Strengthens UN coordination in Colombia by transforming fragmented project data across agencies into a verified, searchable knowledge base, enabling the Resident Coordinator Office to respond to territorial queries faster and support more evidence-based decision-making. Improves data governance and cross-agency collaboration through AI-powered tools that consolidate information, validate inputs, and make selected insights accessible for wider use
Progress so far
Random Monkey has build a data governance and knowledge platform for the Resident Coordinator Office in Colombia that centralizes fragmented UN project information across agencies and enables real-time territorial queries in 2025.
Workplan
Random Monkey will adapt its existing retrieval-augmented generation architecture, develop a centralized data lake, pilot a territorial chatbot and validation system using priority 4W data, and align the information architecture with the needs of the RCO/OCR and UN Country Team. The solution will support faster, evidence-based coordination across approximately 25 UN agencies, with selected insights and guidance made available through a public-facing knowledge base for wider consultation.
Project Snapshot
The project would support the UN Resident Coordinator Office in Colombia by strengthening its knowledge base of UN projects across the country. It addresses a current challenge where data management across approximately 25 UN agencies remains fragmented and manual, making it difficult to answer territorial queries quickly. The proposed approach would transform this into a more integrated and automated system, enabling real-time responses to questions that currently take days to resolve.
The solution is built around three interconnected components: a centralized data lake, a territorial retrieval-augmented generation chatbot that can answer natural-language questions about UN cooperation projects across the territory, and an intelligent validation system.
The core technology is consistent with Random Monkey’s existing approach: the territorial chatbot uses RAG, so the underlying retrieval-augmented generation architecture is reusable. However, the information architecture and data sources would need to be adapted to the needs and requirements of the UN office and this would constitute part of Random Monkey’s workplan over Aug-Nov 2026.
The current scope prioritizes 4W data, but the project could later incorporate additional sources from other UN systems, particularly staff and operations data. This would provide a fuller operational picture, including where personnel are located and what is spent operationally in each territory, directly supporting the management and coordination of the UN team in Colombia.
The project is currently operating at pilot level and is seeking funding to deploy live across contexts and scale. Its primary beneficiary is the UN coordination system itself: it would equip UN offices and agencies in Colombia to answer territorial queries in real time and coordinate cooperation more effectively. The public-facing knowledge base would also make selected recommendations and guides available for wider consultation.
UN project data in Colombia is currently fragmented across agencies, systems, and reporting processes, making it difficult for the Resident Coordinator Office to quickly access a complete picture of who is doing what, where, and with which resources. This limits the RCO/OCR’s ability to respond to territorial queries in real time, coordinate effectively across approximately 25 UN agencies, and support timely, evidence-based decision-making for the UN Country Team.
Random Monkey will adapt its existing retrieval-augmented generation architecture into an AI-powered data governance and knowledge platform for the RCO/OCR in Colombia. The solution will centralize priority 4W data into a shared data lake, enable natural-language territorial queries through a chatbot, and include a validation system to improve the reliability of information. By consolidating fragmented project data into a verified and searchable knowledge base, the platform will strengthen cross-agency coordination, support faster responses to country-level information needs, and make selected insights available through a public-facing knowledge base.
Colombia