Of the four language skills, speaking, listening, reading, and writing, speaking is by far the most important and the most difficult to develop. Many English language learners (ELLs) are reluctant to practice speaking, due to shyness or fear of making embarrassing mistakes. Spoken language activities in class are time-consuming and difficult to manage when many students are speaking at once.
Alelo Enskill®, an AI-driven learning platform, addresses these problems. Enskill helps students build self-confidence and make rapid progress toward proficiency. It also reduces teacher workload and makes classroom activities more effective.
Learners access Enskill through a web browser on their computer or mobile device. They converse with interactive characters by speaking into a microphone. The on-screen character interprets the learner’s speech and responds, and at the same time evaluates the learner’s communication skills. Simulations are aligned with proficiency levels (A1, A2, etc.) on the Common European Framework of Reference (CEFR) scale. Thanks to Enskill’s advanced natural language processing technology, learners can express themselves in a variety of ways and are not confined to a fixed script. At the end of each conversation Enskill provides feedback, including quantitative metrics of performance, and recommends exercises for further practice.
UVM (Universidad del Valle de México) Toluca Campus has integrated Alelo Enskill into its English curriculum. At the beginning of each week the instructor goes over new language forms and structures in class. Afterwards they can practice these forms and structures in the Enskill simulations. They can practice on computers in the computer lab, and continue to practice at home as needed. Afterwards learners participate in a speaking practice class, in which they apply the speaking skills that they have learned.
We analyzed the performance data collected by Enskill from learners participating in the UVM program. Enskill saves anonymized speech recordings and other learner data in the cloud, making archival analysis possible.
UVM learners practiced Enskill conversational simulations repeatedly, and their performance improved with practice. On average the CEFR A1 users who practiced simulations multiple times achieved an increase of 10.97% in mastery score per trial. The A2 users who practiced simulations multiple times achieved an increase of 14.15% in mastery score per trial. Students reported that Enskill helped them improve their English, and teachers reported that students came to class better prepared and ready to engage.
The analysis also showed that learners retained their spoken language skills well over time. Comparison of student performance before and after gaps in practice show that performance degrades very slowly, particularly for students at higher levels of language proficiency.
This study demonstrates the potential of cloud-based data-driven learning platforms to transform learning and assessment. Data collected by Alelo Enskill is used on an ongoing basis to monitor and assess learning, retrain AI models, and system inform development. Real-time formative assessments make possible immediate feedback and personalized learning, and track each learner’s progress toward proficiency.
Click here to download the full report.