The End-Year Summary Narrative Quality Assurance Tool is a web-based tool developed by inmind.ai for UNICEF. It aims to provide two measures of quality assurance for the end-of-year summary reports.
Those reports are created by each country’s UNICEF office to summarize the work done throughout the year by that country and are submitted to UNICEF on a yearly basis. Their structure is defined each year by the main UNICEF office who puts a number of sections with a specific topic for each section.
The tool developed by inmind .ai enables country offices to upload their reports and using a machine learning model they will be able to determine if the sections of their reports are compliant to the UNICEF guidelines. This is called an “adherence to guidelines” check.
In addition, another type of checks called “reputational risk” will take place upon upload. This check uses a different machine learning model which flags sensitive or offensive sentences based on a set of predefined sensitive categories. Finally, the app leverages the power of regular expressions to ensure the adherence of documents to UNICEF's editorial guidelines
