evidoria

← Back to browse

Good practice Imported

ICFES Cuts National Exam Grading Time 91.6% with AWS/Blend AI-Assisted Coding of Open Responses

Colombia · Bogota · See the Colombia profile

Colombia's national exam body ICFES, working with AWS and Blend since 2022, automated coding of open-ended Saber-exam responses - cutting grading time 91.6% and results-delivery time about 50% - and used Amazon Rekognition to verify identity and proctor over 383,000 virtual Saber

ICFES Cuts National Exam Grading Time 91.6% with AWS/Blend AI-Assisted Coding of Open Responses

Details

Promoter
Instituto Colombiano para la Evaluacion de la Educacion (ICFES)
Period
2017-2025 (AWS partnership since 2017; AI-grading transformation since 2022, announced Sept 2025)
Keywords
national examinations, automated grading, virtual proctoring, public-sector AI

Description

The Instituto Colombiano para la Evaluacion de la Educacion (ICFES), Colombia's national standardised-testing body, announced on 4 September 2025 a technological transformation of its evaluation system built with AWS, its cloud partner since 2017, and Blend, an AI/data-science provider. The centrepiece is an AI accelerator that automatically codes open-ended responses on the Saber 11, Saber Pro and TyT Saber Pro exams, which ICFES and Blend say reduces grading time by 91.6% and cuts the time to deliver results by roughly 50%.

During the COVID-19 pandemic, ICFES used Amazon Rekognition image- and video-analysis to verify candidate identity and proctor exams remotely, processing more than 383,547 virtual TyT Saber Pro tests. The system also runs geolocation algorithms that assign candidates to their nearest exam site while accounting for geographic barriers such as rivers, and stores exam data in a 'Data Lake' alongside generative-AI report generators intended to give schools and policymakers faster, more personalised results.

ICFES director Elizabeth Blandon Bermudez said the aim was 'reducing times, optimising resources and offering students... more reliable tools,' while emphasising a 'human and collaborative component' alongside the automation; Blend regional leader Andres Barrantes and AWS's Karla Wong were quoted framing the partnership as public-sector AI with 'real impact.'

The reported figures (91.6% and 383,547) come from the implementing companies' and ICFES's own announcements rather than an independent audit, and neither the September 2025 announcement nor contemporaneous press coverage describes a human-review step for the automated open-response coding itself, only for identity verification and proctoring - an important gap given the scale and high stakes of a national university-entrance and completion exam.

Read the full analysis: https://www.icfes.gov.co/comunicado-de-prensa-icfes-impulsa-la-transformacion-tecnologica-del-sistema-de-evaluacion-del-pais/

Implementation

Implementation detail (cost, timeline, staffing, conditions for success) is not yet available for this practice.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful