Analyse classroom discussions with AI
Review anonymised group responses, identify recurring misconceptions and keep contrary evidence in view when planning the next lesson.
Analyse classroom discussions with AI by asking what the class appears to understand and what evidence might challenge that reading. A group may use the right vocabulary while explaining a concept in two incompatible ways. The discussion-insights-for-education skill reports recurring patterns at cohort level.
Set the time window and use approved, de-identified material. A discussion and a written exercise may suggest different conclusions. The workflow preserves the conflicting sources and asks the teacher which one should govern the current record. A later comparison can show a shift without attributing it automatically to a particular lesson.
What it does
- Group similar questions and estimate how often they appeared.
- Remove identifying detail from illustrative phrasing.
- Retain responses that challenge a tentative interpretation.
- Compare an earlier cohort pattern with a later cycle.
- Choose a teaching response that addresses a specific misconception.
Install and use
- Use an authorised Microsoft 365 Copilot environment with metered Cowork access and the Education plugin. Define the cohort, cycle and permitted sources.
- Ask, “Compare these anonymised explanations of percentages. Show the common interpretation, contrary examples and a possible follow-up exercise.”
- Review the findings and decide how to handle any conflicting sources before recording the agreed next step.
FAQ
Source
Made by Microsoft Copilot Cowork team. Original skill: discussion-insights-for-education, published on GitHub under the MIT licence.
View on GitHub