Engagement falls as school goes on
Reported student engagement drops from 76% in grade 5 to 44% in grade 12.
Gallup student poll, as reported by Education Evolving.
Any AI can answer a question. CourseBend is built to know who’s asking: what they love, where they’re headed, and what their school requires. It’s augmentive by design: it sits on top of the courses a student already has and adds to what they can do. Nothing about the requirement moves. Only the way in.
Pull the requirement from the school of record and the published standard. Cite it. Never generate it from memory.
Work out why this topic matters to this student, and let their interest shape the problem itself, not just decorate it.
Lead with context. MalachAI explains, asks and guides, and never does the student’s thinking for them.
Ask for it back in their own words. Parroting doesn’t count. The grade stays with the school.
File each observation into the student’s record the moment it happens, so the next lesson starts from what’s true.
Putting dinosaurs in a word problem is flavouring, and anyone can do it. A bend restructures the way in. A student who wants to design snowboards can meet 1800s wheat disputes through materials, tariffs and supply chains. Same outcome. Very different reason to care.
| Works on the school’s real assignments | Works on something else | |
|---|---|---|
| Knows the student | CourseBend is built to do both | Personal, but the required work stays untouched |
| Doesn’t know the student | Helps with the assignment, for anyone at all | General lessons and practice |
Math, a parent’s real-estate lesson, a boxing obsession and a grade 12 Spanish course aren’t separate files. Every observation, whether it’s prescribed, pushed or pulled by curiosity, lands in one connected record.
Level 5 is what earns a badge. And it assumes you forget: after a gap, MalachAI refreshes at the level a student last reached, and they prove it again. Old knowledge is a locked box. The refresher is the key.
Learning takes hold when it connects to what a student cares about, and when the student still does the work. CourseBend is built on that research.
Reported student engagement drops from 76% in grade 5 to 44% in grade 12.
Gallup student poll, as reported by Education Evolving.
About 1,000 high school students used GPT-4. Those with open access did 17% worse once it was taken away. A version with guardrails largely avoided the harm. Guardrails are the point.
Bastani et al., PNAS (2025).
In randomized trials, a short exercise linking a lesson to a student’s life helped those struggling most. Ninth-grade science students who expected to do poorly gained nearly two-thirds of a letter grade over controls.
Hulleman & Harackiewicz, Science (2009); Hulleman et al., J. Educational Psychology (2017).
Ninth graders in algebra who got problems matched to their out-of-school interests solved problems better and worked more efficiently. CourseBend carries that idea through every required course, across all of high school.
Walkington (2013); Bernacki & Walkington (2018).
The passion bend and the four-question demo. Onboarding work with BC students.
The first version: MalachAI on phone and desktop, snapshots of paper work, and a basic road. First pilot in the new year.
Parent and teacher portals, the full record view, and everything on the Road.