Comparisons
Where we lose, written down.
A comparison in which one product wins every row is an advertisement, and a reader can tell. Each page below names at least one thing the other product does better, and none of them quote a competitor’s price — we cannot verify it and a stale number is the one claim you can check.
- Zelvi vs QuizletA shared library, or your own course material.These are not the same product competing on features. One gives you what other students have already made; the other turns what your instructor published into practice.Read the comparison
- Zelvi vs ChatGPTGeneral knowledge, or your course.ChatGPT is a better general assistant than anything we will build. It is not grounded in your syllabus, and for coursework that is the property that decides the outcome.Read the comparison
- Zelvi vs AnkiThe best repetition engine, or the material to put in it.Anki solves retention and asks you to author everything. Zelvi authors from your documents and teaches the concept. Many people should use both.Read the comparison
- Zelvi vs SchoolAITeacher-designed Spaces, or a recovery loop grounded in the course.SchoolAI is built around the K-12 teacher and the classroom. Zelvi is built around what happens after a student gets a question wrong in a course with published material, and the record that leaves behind.Read the comparison
- Zelvi vs MagicSchoolEighty tools, or one loop that follows through.MagicSchool saves teachers time across a very wide range of tasks. Zelvi does fewer things, all from the course’s own material, and records whether a struggling student recovered.Read the comparison
- Zelvi vs BriskA tool in your browser, or a tutor in your course.Brisk helps the teacher inside Docs, PDFs and web pages. Zelvi works with the student, from the course’s published material, and hands the teacher a record of who recovered.Read the comparison
- Zelvi vs Kahoot!A great game, or a game that follows through.Kahoot! is the better show. Zelvi is built for the minute after the last question: every miss becomes that student’s practice, and the instructor gets what to reteach.Read the comparison
The institutional AI-tutoring landscape
Every category has a strong product. None of them is the same product as Zelvi.
Five categories a dean is likely to have on the shortlist, and the specific thing Zelvi does differently in each. Written to be fair to the other product; a comparison in which we win every row is an advertisement.
| Category | What they do | What Zelvi does differently |
|---|---|---|
| Generic AI assistantsChatGPT, Gemini for Education, Claude | General chatbot, no course scope, no grounding you can measure. | Refuses to answer what is not in the material, and the dean can measure the refusal rate on the cohort dashboard. |
| AI tutors on someone else’s contentKhanmigo (Khan Academy) | Grounded in Khan’s own content library only. | Grounded in the institution’s own material — no vendor lock-in on curriculum, and the same tutor runs on the syllabus the faculty already published. |
| Adaptive practice enginesSquirrel AI, Riiid, ALEKS | Adaptive practice from a proprietary model whose scoring is opaque. | Rubric-based scoring reviewed by course staff; every criterion carries a reference back to the document that authorised it. |
| Platform-native AI coachesCoursera Coach, Duolingo Max | LLM wrapper on top of the platform’s own content. | Installs on top of the institution’s content and the cohort dashboard measures the outcome the institution already reports on. |
| Cohort-analytics tutorsCentury Tech, Cognii | Cohort dashboards focused on engagement time. | The dashboard reports recovery rate on stuck moments, not minutes-on-task, and the escalation queue routes an unresolved learner to a real human on faculty. |
Why the four together
The moat is the combination: grounded refusal · rendered diagram · measurable recovery rate · escalation to a human.
Any one of those alone is not defensible — a competitor can add a citation, license a diagram library, ship a dashboard, or wire a support queue on any given quarter. The four wired to the same attempt chain, inside the institution’s own material, is the shape that takes a real pilot to build.