When you export tutor insights for a classroom, every student gets a row with the same set of columns. Here is what each one means and how to read it.
Time columns
Total Time The full elapsed time the student had a conversation open, including pauses where they may have stepped away or were thinking. This is the broadest measure of how long the conversation was alive.
Active Time The time the student was actively in a conversation, ignoring lengthy pauses or gaps. This is the time genuinely engaged with the tutor, with the idle time stripped out. If a student opens a tutor for an hour but only sends messages for 10 minutes, Active Time is 10 minutes.
Engagement columns
Questions Asked How many messages the student sent to the tutor.
Questions Answered How many of the tutor's questions the student responded to. The tutor will often ask follow-up questions to push the student's thinking; this counts the ones the student actually answered.
Answer Quality
A four-level classification of how the student is engaging with the tutor.
Poor Treats it like Google or ChatGPT. Throws questions at the tutor and does not engage. Copy-pastes problems and expects answers. Gives up quickly when the tutor pushes back.
Satisfactory Engages, but surface-level. Gives short answers. Accepts the tutor's hints without thinking them through. Does enough to move on, but does not stretch.
Good Genuine attempts. Builds on what the tutor says. Explains their thinking even when not asked to. Asks clarifying questions of their own when stuck.
Excellent Thoughtful engagement throughout. Makes connections across concepts. Pushes back on ideas, asks probing questions, and demonstrates understanding through reasoning rather than waiting to be told. Often goes beyond what the lesson required.
Group
A three-level grouping of where the student is in their grasp of the material.
Beginner Needs frequent scaffolding. Struggles with foundational concepts in this topic. Often asks the tutor to explain things multiple times. Benefits from being pointed back to core definitions and worked examples.
Intermediate Has the basics but needs guidance on application. Can work through problems with prompting. Sometimes makes connections without being led there, but not consistently.
Advanced Confident with the material. Working at higher-order thinking - applying concepts to new contexts, comparing methods, or extending ideas. Ready to be stretched.
Misconception columns
These columns surface specific gaps the tutor has identified through conversation. They are populated when the tutor sees a recurring or load-bearing misunderstanding, not for every minor wobble.
Misconceptions: Concepts The specific concepts the student is misunderstanding.
Misconceptions: Details What exactly the student is getting wrong about each concept.
Misconceptions: Syllabus Connection Which part of the curriculum or learning outcome this misconception affects.
Misconceptions: Root Causes What is likely driving the misconception. Often this points to a foundational gap rather than the surface-level error.
Misconceptions: Worth Reteaching A flag for whether this is a misconception worth addressing in class. Useful for prioritising re-teach moments when you have many students with many small gaps.
How to use these columns together
Answer Quality and Group are the two columns most useful for differentiating follow-up:
Beginner / Poor: priority for a re-teach. They are stuck and not yet getting unstuck on their own.
Beginner / Good or Excellent: working hard but on shaky foundations. Worth a short check-in to make sure the gap closes.
Intermediate / Satisfactory: the silent middle. Often missed in class. The insights flag them.
Advanced / Excellent: ready for stretch tasks or peer-teaching opportunities.
The misconception columns then tell you what to actually do with each group.
Things to know
Insights are generated from the conversation itself. A student who has not had a meaningful conversation yet will appear without classifications.
Active Time and Questions Asked are mechanical counts. Answer Quality, Group, and the misconception columns are interpreted from the conversation content.
Classifications can shift as a student has more conversations. A student classified as Beginner / Poor in week one might be Intermediate / Good by week three. The export reflects their most recent state.