Document Processing · Real-time Analytics
Lithuania’s National Agency for Education needed to extract, standardize, and compare information from thousands of mixed-format school management reports because manual extraction took hours per report and made nationwide evidence-based education planning impractical.
Reported impact
NAE eliminated almost all manual extraction work, gave leaders a much faster way to compare schools and municipalities, and created a data foundation for longer-term education policy and performance analysis.
The National Agency for Education worked with IBM Business Partner S2P to build an AI-enriched document-ingestion and analytics pipeline on IBM watsonx.ai and IBM Cognos Analytics. Each year the agency receives around 3,000 school management reports, averaging about ten pages and combining structured and unstructured content across many file formats. S2P designed a process that first extracts and transforms the raw report content into a vector database, then uses watsonx.ai to score each report against 14 client-defined key performance criteria on a zero-to-three scale. Once that information is structured, it is blended with other data from the national Education Management System, which itself aggregates information from more than 40 registers. IBM Cognos Analytics then turns the combined dataset into executive dashboards so agency leaders can compare schools, municipalities, and national trends. The outcome is a document-to-policy workflow that replaces manual extraction with repeatable AI-assisted analysis and gives education leaders a longitudinal, data-driven view of school performance.
Mixed-format school reports are ingested and transformed into a vector database, watsonx.ai evaluates the content against 14 KPIs, the resulting scores are combined with structured data from the Education Management System’s 40-plus registers, and Cognos dashboards present the full picture to agency leaders.
“99% reduction in manual labor”
“Every year, around 3,000 new reports are submitted”
“Initially, the raw data is extracted, transformed and inserted into a vector database.”
“Then, the AI evaluates the information and delivers a figure for each KPI on a scale of zero to three”
“Working with IBM Business Partner S2P, the NAE chose IBM watsonx.ai and IBM Cognos Analytics”
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
IBM Sample Apr 4, 2025 New Jersey Department of Community Affairs The New Jersey Department of Community Affairs needed to replace fragmented, paper-based benefits administration across multiple agencies with a unified way to gather applications, review documentation, adjudicate… 60% reduction in average time for a resident to complete a USFHEA application IBM Consulting Advantage for Cloud Transformation · Microsoft Azure AI Document Intelligence · Microsoft Azure Databricks 
Microsoft Sample Aug 14, 2024 Georgia Tech Dr. Omar Asensio, an Associate Professor at Georgia Tech and Visiting Fellow at Harvard Business School, leverages Microsoft's Azure OpenAI Service to enhance the electric vehicle (EV) charging infrastructure. By… A decade ago, electric vehicle sales barely accounted for 1% of the total new vehicles sold. Fast forward to… Azure OpenAI Service IBM Sample Sep 15, 2025 The University of Auckland University of Auckland researchers were spending too much time navigating grant, ethics, and export-control requirements because funding applications and compliance reviews involved multiple forms, committees, and rules… 85% accuracy in detecting known error types in sample applications IBM Cloud · IBM watsonx Assistant · IBM watsonx.ai 