Teach the AI to be you

The difference between generic AI output and output that sounds like it came from you is context — and TeachersFlow is built to hold that context permanently, so you provide it once instead of re-explaining yourself in every prompt.

The layers

All of this lives in Account settings -> Personalization, and everything below is applied automatically once set:

  1. Teaching profile — who you are as a teacher: subjects, level, philosophy, what you value in student work. Injected into every AI generation across the app. One honest paragraph beats a long generic one.
  2. Per-tool defaults — your standing instructions per tool: default grading instructions and point scales for the Test grader, default requirements for the Assessment generator, default lesson requirements for the Lesson planner. Anything you'd otherwise retype every time belongs here.
  3. Training documents — real examples of your work. Upload .docx files or Google Docs, or import directly from your history, teaching materials, and groups. Variety beats volume: different levels, topics, and styles teach your voice better than ten similar documents.
  4. Ratings — rate generated assessments (1–5 stars). Over a few weeks this teaches the AI not just your style but your quality bar.

Why this compounds

Each layer narrows the gap between "an AI wrote this" and "I wrote this on a good day." Combined with teaching goals and student history, it's why TeachersFlow output gets more yours over time — the app accumulates, in structure, what a chat tool forgets the moment the conversation ends.

Practical order

If you're setting up fresh: write the teaching profile first (five minutes, biggest single effect), add per-tool defaults the first time you catch yourself retyping something, and import training documents once you've produced or exported a few pieces you're happy with. Deeper tips: Pro tips.