Enterprise Knowledge Search
UNESCO wanted to make large-scale learning-assessment data far easier for policymakers and educators to analyze so they could make better decisions about foundational reading outcomes.
Reported impact
UNESCO gained a working agentic-AI approach for turning large education datasets into accessible policy insight, with support for open-source reuse and broader evidence-based decision-making across countries.
UNESCO worked with Capgemini and AWS to build an agentic AI education-policy assistant around the PIRLS 2021 reading-literacy dataset. The project used the Global Data Science Challenge as a development engine: nearly 1,500 participants created and tested large-language-model systems that could analyze structured assessment data and answer detailed policy questions about learning trends, student performance, and teaching practices. The winning insAIghtED architecture combines multiple agents—a lead analyst, data engineer, chart preparer, and data scientist—to query PIRLS data, pull in supporting sources such as GDP, population, and life-expectancy data, and return visualizations plus narrative explanations. Amazon Bedrock and SageMaker supported model development, while LangChain and crewAI coordinated multi-agent reasoning so policymakers can move from raw data to evidence-backed recommendations more quickly.
PIRLS 2021 assessment data and supplemental public datasets such as GDP, population, and life expectancy are queried by coordinated agents that aggregate, analyze, and visualize the information. Amazon Bedrock and Amazon SageMaker provide the model layer, while LangChain and crewAI orchestrate reasoning so the system can return charts, short answers, and fuller explanations for policy users.
"Across Capgemini, teams used AWS technology, including Amazon Bedrock and SageMaker"
"By integrating various data sources and employing state-of-the-art AI frameworks like Langchain and crewAI, the insAIghtED model delivers impactful, real-world insights"
"The architecture of the solution involves a lead data analyst, data engineer, chart preparer, and data scientist"
"participants were tasked with building an AI education policy expert capable of analyzing and providing insights from the PIRLS 2021 dataset"
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Microsoft Sample Oct 15, 2026 Bar-Ilan University Students who miss lectures struggle to reconnect new lessons with prior material when course information is spread across notes, recordings, slides, and readings, and lecturers need a reliable way to answer routine… 71 courses Azure AI Services · Microsoft Azure · Microsoft Copilot Studio 
NVIDIA Sample Aug 3, 2026 StudyFetch StudyFetch needed to keep AI tutoring affordable while scaling live lecture transcription and personalized learning for millions of students, but off-the-shelf cloud transcription created cost and latency ceilings that… 100M+ learning interactions NVIDIA H100 GPUs · NVIDIA L40S GPUs · NVIDIA NIM NVIDIA Sample Apr 16, 2026 University of Florida The University of Florida wanted a campus-wide AI platform that could serve students, faculty, staff, and researchers with many models and assistants while controlling cost, preserving data privacy, and keeping… 1,500+ unique uses of NaviGator ToolKit HiPerGator · Llama NIM · NVIDIA NIM microservices 
