AI video assessment and feedback assistant for ESOL tests

ESOL (English for Speakers of Other Languages) speaking and listening assessments are currently recorded and manually reviewed by teachers, making the process time consuming, labour intensive and inconsistent. As a result, assessments are not always carried out regularly, and learners often wait too long for feedback. This delay limits their ability to improve key language skills, affecting confidence and progression into further education or employment. With over 160,000 ESOL learners in England, there is clear demand for more timely, effective feedback, which evidence shows is critical in supporting learner progression and access to opportunities.

This project will develop and pilot an AI-powered video assessment tool for ESOL speaking and listening. The tool will analyse recorded assessment videos using teacher-defined mark schemes, generating structured feedback aligned to ESOL criteria. Teachers remain in control, reviewing and refining the AI outputs before sharing with learners. Through a chatbot interface, learners will receive on-demand feedback and will be able to ask questions to deepen their understanding. The platform will also generate insights across groups of learners to support moderation and highlight common skills gaps.

This approach could significantly reduce teacher workload while improving the speed and consistency of assessment feedback, helping learners improve confidence and language skills faster. Drawing on Second Mesh’s expertise in further education and AI, this approach is well placed to support more consistent ESOL delivery. At scale, the solution could be adapted for other subjects and assessment contexts across the FE sector.