Muse for Educators: Lesson Plans, Differentiated Materials, and Grading Drafts That Start From Your Standards
An educator's working setup for Muse: a Sunday planning desk that drafts next week's lessons from your actual standards, adapts one lesson three ways, and…
By Professor Muse11 min read
Teaching has two jobs, and only one of them happens in front of students. The other happens at a kitchen table on Sunday night: turning a standard into a lesson, turning that lesson into versions for students who read at different levels, turning a stack of exit tickets into feedback that is actually useful. That second job is where the week gets won or lost, and it is the job Muse is good at in an educator’s seat.
The evidence for the time saving is real but modest, and worth stating plainly. In a randomised trial run by the National Foundation for Educational Research and published by the Education Endowment Foundation in October 2026, primary teachers using an AI lesson-planning assistant spent an average of two and a half hours a week on lesson planning and resource preparation, against three hours and 19 minutes for teachers who did not use the tool, a saving of around 49 minutes a week. The same trial found that use of the tool declined over ten weeks, and fewer than one in five teachers felt it meaningfully changed their planning. An AI draft saves time. The teacher’s judgment about what a class needs stays with the teacher, and the teachers in the trial who got the most from it were the ones who treated the draft as a starting point.
That is the rule for this lesson: Muse drafts, you teach. Every plan, every differentiated version, and every piece of feedback passes through your eyes before it reaches a student. Muse never sees a student’s name, grade, or personal information, and it never makes a grading decision. It prepares the work around teaching. You keep the work of teaching.
One boundary first, because it shapes the whole setup. Muse requires users to be 18 or older under its terms, so this lesson is about the teacher’s own workflow, not about handing Muse to students. Students never touch this setup. What you build here lives on your side of the desk.
The short version
- Run three jobs on a rhythm: a Sunday lesson-plan draft for the coming week, differentiated versions of each lesson, and rubric-based feedback drafts on short student work.
- Paste the actual standard into every prompt. A plan built from a standard you quoted is checkable. A plan built from a standard Muse recalled from memory is a guess dressed as alignment.
- Keep student names, grades, and personal details out of every prompt. Anonymize the work before it goes in, and reattach names only after the draft comes back.
- The failure mode to watch for is a polished plan that teaches the wrong standard. The fix is a two-minute alignment check against the standard text, every time.
The setup: The Sunday Desk
Call the whole arrangement the Sunday Desk. I run it as a seventh-grade science teacher with three sections of the same course, one planning period a day, and a unit on ecosystems coming up. The desk has three jobs, and each one maps to a prompt below.
For the weekly plan draft, every Sunday Muse drafts next week’s lessons from the standards and the materials you already have: the textbook chapter, last year’s plan, and the lab supplies on hand. Each draft has an objective, a timed sequence, the materials needed, and a check for understanding at the end. You edit it into the lesson you will teach.
For the differentiated versions, Muse produces three versions of the same core activity for each lesson: a supported version with sentence frames and a worked example, a standard version, and an extension version with a harder question and less scaffolding. The objective and content stay the same; only the amount of support changes. You pick which students get which version, because that call belongs to the teacher who knows the class.
The grading drafts. For short student work, exit tickets and brief written responses, Muse drafts feedback against your rubric: what the response got right, what is missing, and one concrete next step. You read, correct, and approve every draft before it goes back. You return nothing to a student until you have read it.
| Job | Muse does | You do |
|---|---|---|
| Weekly plan draft | Drafts lessons from your pasted standards, textbook chapter, and last year’s plan | Edits the draft into the lesson you teach, checks alignment, adjusts timing |
| Differentiated versions | Produces supported, standard, and extension versions of the same activity | Decides which students get which version, fixes anything that misses the class |
| Grading drafts | Drafts rubric-based feedback on anonymized short student work | Reads and corrects every draft, makes the grading call, returns it yourself |
| Parent communications | Drafts a weekly update or a conference note from your bullet points | Rewrites in your voice, checks every fact, sends it yourself |
Before any of this runs, set the standing instructions from our memory guide: your subject and grade level, your lesson-plan format, the sentence frames your students already know, and the rule that student names and grades never go into a prompt. That instruction, stored once, is what keeps every Sunday draft in the same shape. The Sunday run itself is a scheduled routine, the same shape as a morning briefing, with one extra rule: if the standard text is missing from the prompt, Muse says so and stops instead of drafting from memory.
The Sunday Desk builds on our guide to writing briefs Muse can actually use:
Prompt 1: the weekly lesson-plan draft
Run this every Sunday, one prompt per lesson, with the standard pasted in full.
Draft a 50-minute lesson plan for seventh-grade science on [topic], for the week of [date]. The standard, pasted verbatim, is: [paste the full standard text]. Use only these materials: [textbook chapter or pages], [last year’s plan for this lesson, if you have one], and [lab supplies or resources actually available]. Structure it as: objective in student-friendly language, a timed sequence of activities adding up to 50 minutes, materials needed, and one check for understanding at the end. Do not add standards, topics, or vocabulary that are not in the pasted standard or the listed materials. If a material is missing, say so instead of substituting one. Draft only. I edit it before teaching.
The pasted standard is the whole trick. A prompt that says “a lesson on ecosystems” invites Muse to pick a standard for you, and it will pick one confidently, whether or not it matches the one your district tests. A prompt with the standard pasted in gives you something you can check in two minutes: read the objective, read the standard, confirm they match. That check is the difference between a draft that saves you an hour and a draft that costs you a class period teaching the wrong thing.
Prompt 2: three versions of the same lesson
Run this after you have approved the lesson plan, once per lesson that needs differentiation.
Take this approved lesson plan: [paste the plan]. Produce three versions of the core activity, all teaching the same objective. Version A is the supported version: add sentence frames, a worked example, and break multi-step directions into single steps. Version B is the standard version as written. Version C is the extension version: remove the scaffolding, add one harder question that requires applying the concept to a new situation, and keep the same objective. Do not change the objective, the vocabulary list, or the check for understanding. Label which parts you changed in each version. Draft only. I decide which students get which version.
Keeping the same objective across all three versions keeps differentiation honest. The moment Version A teaches a simpler idea instead of the same idea with more support, you have built three different lessons and graded them as if they were one. The prompt guards against that by naming what may change (support level) and what may not (the objective and the vocabulary). After it runs, read Version A and Version C side by side and confirm they ask for the same understanding.
Prompt 3: feedback drafts on short student work
Run this after collecting exit tickets or short responses, with all student work anonymized first.
Draft feedback on these anonymized student responses to [prompt or question], against this rubric: [paste the rubric]. The responses are labeled Student 1, Student 2, and so on, with no names attached. For each response, write: what the response got right, what is missing or incorrect against the rubric, and one concrete next step the student can take. Keep each feedback note to three sentences. Do not assign a score, a grade, or a level. Do not guess which student wrote which response. If a response is unclear, say so instead of interpreting it generously. Draft only. I read, correct, and approve every note before it goes back.
Three sentences is the limit. Longer feedback drafts drift into generic praise, and generic praise on a wrong answer is worse than no feedback at all. The no-score rule protects that speed. The moment a draft assigns a grade, the teacher has to audit a decision instead of editing a note, and the audit is slower than writing the note from scratch. Keep Muse on the language, and keep the judgment, including the grade, with you.
The parent communication in the table above runs on the same pattern: you give Muse bullet points, it drafts the update or the conference note, and you rewrite it in your voice and check every fact before sending. No student information goes into that prompt either, beyond what you have already decided belongs in a message home, and you review everything before sending it. Our failure modes guide explains why that review step stays mandatory: fluency does not prove the facts inside a draft are right.
What can go wrong: the polished plan that teaches the wrong standard
Here is the failure mode you will meet first. The Sunday drafts come back clean, with objectives, timed activities, materials, and checks for understanding, all in your format. Three weeks in, you notice that one week’s plan covered a standard your course does not test until spring, and skipped the standard the unit was supposed to hit. The plan was well written but wrong, and the wrongness was invisible because the plan looked finished. A draft that looks done invites less checking than a draft that looks rough, which is why it needs more checking.
The fix has two parts. First, Prompt 1 already requires the pasted standard, and you hold it to that: send back any draft whose objective you cannot trace to a line in the standard text. Second, add one standing rule to the desk: before you teach a drafted lesson, read the objective and the check for understanding against the standard, out loud if you have to, and confirm the check tests the objective. That two-minute check catches the drift that a full read-through misses, because the drift hides in the gap between what the lesson claims to teach and what its final question asks.
The same discipline applies to the feedback drafts. A note that praises a response for a point the rubric does not award is a small error with a large cost: the student walks away believing they have mastered something they have not. Read every note against the rubric; a note can sound good and still miss the rubric.
Start with one week
Do not build the whole Sunday Desk this week. Pick next week’s lessons, run Prompt 1 for each of them on Sunday, and read the drafts the way a skeptical editor would: standard in one hand, objective in the other, checking that they match before you touch anything else. If the drafts survive that reading, run Prompt 2 for the lesson that most needs differentiation, then try Prompt 3 on one set of exit tickets, anonymized, with your rubric pasted in.
Run it that way and Muse does what it is good at in an educator’s seat: the first draft, the three versions, the feedback language. You keep what was always yours: which standard the class is actually learning, which students need which version, and what feedback a student deserves. That gives you plans you can defend, materials that fit the class you have, and feedback you have read. The edge is a draft with the standard pasted in, every time.
