Turn teacher data
into AI prompts that work.

A different approach to AI prompts for teachers.

AI prompts for teachers are pre-built questions that run over a school’s own recordings, notes, and survey data: what students think of a unit, what parents raised at conferences, what came out of a professional development session. Curriculum developers, instructional coaches, and school leaders have used a version of this for years, sitting down with a spreadsheet after every survey. The idea worked. The manual version rarely lasted past the first busy month.

Why the ten-prompts list falls apart

A list of prompts is useless without the data behind it, and that’s where most teachers get stuck. Parent-teacher notes live in one binder, exit surveys in a spreadsheet, focus group recordings on a shared drive nobody opens twice. Copying the right prompt into a chat window and pasting in this term’s comments is a task, not a workflow, so it happens once at the end of the year instead of every week when it could still change something.

Reading the data, not just the words

Speak AI treats a stack of survey responses, meeting notes, and classroom recordings the way a full curriculum team would, at machine speed. Text is read for theme and sentiment, and any voice recording, a parent-teacher meeting, an interview with a department head, is read for tone, energy, and emotion, so a comment that reads calm on paper but sounded frustrated in the room isn’t missed. Ask across a term’s worth of notes with AI chat, using the same prompt workflows teachers once ran one file at a time, now running natively over the whole folder with ChatGPT, Claude, and Gemini built in, or over the API and MCP if your district already builds its own tools.

Ten prompts, one workflow

• Curriculum effectiveness: “What are students saying about this unit, and where should we adjust it before the next one starts?”
• Parent-teacher communication: “What did families raise at this round of conferences, and what still needs a follow-up?”
• Student engagement: “Which parts of the lesson lost students, and which brought them back in?”
• Professional development: “What did peer observers actually flag, the praise and the concerns?”
• Qualitative research on teaching methods: “What do the last two years of classroom research say about this method?”
• Ed-tech pilot feedback: “Is this new classroom tool worth rolling out district-wide, based on what teachers and students actually said?”
• Course & survey adjustment: “Where did satisfaction drop this term, and what specifically caused it?”
• Classroom recordings: “What does the recording show about pacing, participation, and tone that a lesson plan can’t?”
• Planning meetings & expert interviews: “What did we decide, who owns it, and what did the visiting specialist actually recommend?”

The result is less a stack of prompts and more a standing habit: a term’s worth of notes, surveys, and recordings answered in minutes instead of read once and filed away. Trends across classes and semesters become dashboards you can customize and white-label for a department chair or principal, instead of a hunch repeated at the next staff meeting. The same underlying system, at a very different scale, is what let a legal intelligence firm run this kind of analysis over 5,100+ hours of calls and save $700K doing it (large-scale comparative analysis); a department running it over a term of feedback works the same way, just smaller. And because classroom data rarely lives alone, the same engine also handles call scoring for admissions and intake calls on the same platform.