Copied to clipboard!
Module 4 For Educators intermediate 50 min

Planning with AI

From a real goal to a structured plan to evidence it worked.

For Teachers · Lessons, courses, feedback, family comms

You're viewing the Administrator track. These Teacher exercises are also here. Switch to Teacher in the top bar to focus them.

Build a Lesson or Unit Plan with AI

A teacher needs a lesson or short unit for a specific standard and a specific class, with the right hook, scaffolding, checks for understanding, and differentiation for the actual students in the room (English learners, students with IEPs or 504s, a range of reading levels). The teacher wants AI to do the structural heavy lifting (drafting the arc, the timing, the materials) while they keep the pedagogical judgment about what their learners actually need.

Recommended tool

claude

Claude is strong at structured, multi-part planning: turning a standard plus your real class constraints into a coherent lesson with a clear arc, checks for understanding, and differentiation for the students in front of you. If the lesson must be anchored to an uploaded curriculum or textbook PDF, NotebookLM is the better choice because it keeps the AI grounded in that one document.

Bring to the exercise

  • The standard(s) or learning target the lesson must hit
  • Your class context (class size, period length, reading-level range, number of English learners, number of students with IEPs or 504s). Use counts and role labels, not names
  • Any curriculum, textbook, or pacing guide the lesson must align to
  • 2 lessons that worked well + 1 that flopped (so the AI can see what “good” and “bad” look like for you)
  • Your Educator Context Document from Module 2
Step 1

Define the Standard and Your Class

Get AI to understand the target and the room before it plans anything. AI cannot see your students; you have to describe them. Use counts and role labels, never real student names.

What to substitute before pasting

  • [INSERT STANDARD OR LEARNING TARGET] The exact standard code or learning target this lesson must hit.
  • class context fields Counts and ranges only: class size, period length, reading levels, number of ELs, number of IEPs/504s. Never type a student’s name.
  • [PASTE YOUR EDUCATOR CONTEXT DOCUMENT] The context document you built in Module 2: how you teach, your routines, your students.
I am a classroom teacher planning a lesson. Here is the standard or learning target: [INSERT STANDARD OR LEARNING TARGET].

Here is my class context:
- Subject and grade: [SUBJECT / GRADE]
- Class size: [NUMBER OF STUDENTS]
- Period length: [MINUTES PER PERIOD]
- Reading-level range: [LOWEST TO HIGHEST READING LEVEL]
- English learners: [NUMBER] (proficiency levels: [LEVELS])
- Students with IEPs or 504s: [NUMBER] (types of accommodations: [ACCOMMODATION TYPES])

Here is more about how I teach and who my students are:
[PASTE YOUR EDUCATOR CONTEXT DOCUMENT]

Before you draft anything, ask me clarifying questions about the standard, the class, or what I want students to walk away able to do.
What good output looks like
  • Restates the standard in plain language and names what students should be able to do by the end.
  • Asks at least 2 clarifying questions before drafting a lesson.
  • References your class context (size, reading range, ELs, IEPs) instead of assuming a generic class.

Verify before using AI's output

Step 2

Set Objectives and a Hook

Turn the standard into clear, observable objectives and an opening that pulls these specific students in, not a generic attention-grabber.

Using the standard and class context above, do two things.

1. Write 1–3 specific, observable objectives for this lesson. Each one should finish the sentence "By the end, students will be able to…" and should be something I could actually check.

2. Design a hook (the first 5–10 minutes) that:
- Connects directly to this standard
- Fits the age, interests, and reading range of my students
- Gives every student a way in, including my English learners and students who read below grade level

Explain in one line why the hook fits these students, not just any class.
What good output looks like
  • Objectives are observable and tied to the standard, not vague topics like “understand fractions.”
  • The hook connects to the standard and gives lower-readers and English learners a way in.
  • The AI explains why the hook fits your class, not a generic one.

Verify before using AI's output

Step 3

Build the Lesson Arc

Move from objectives to a real teaching sequence, with a check for understanding at each phase so you catch confusion before the bell, not after the test.

Build the lesson arc that moves students from the hook to the objectives. Use one of these structures (pick the one that fits this standard and tell me why):
- I-do / we-do / you-do
- Launch – explore – summarize

For each phase, give me:
- Purpose of the phase
- What I (the teacher) do
- What students do
- A specific check for understanding: how I will know, in that moment, whether students are getting it
- Rough timing in minutes that fits my period length

Make the checks for understanding quick and observable, not "ask if there are questions."
What good output looks like
  • Each phase names a purpose, a teacher move, a student action, and a check for understanding.
  • Checks for understanding are quick and observable (thumbs, exit prompt, quick write), not “any questions?”
  • Timing adds up to your actual period length.

Verify before using AI's output

Step 4

Differentiate for Your Real Roster

Make the lesson reachable for the students actually in the room. Describe needs by group and accommodation type. Use role labels like “Student A,” never real names.

What to substitute before pasting

  • [PROFICIENCY LEVELS] The English-proficiency levels in your class (for example, entering, emerging, expanding).
  • [ACCOMMODATION TYPES] The kinds of accommodations on your students’ IEPs/504s: extended time, read-aloud, chunked tasks. Describe the accommodation, not the student.
Now differentiate this lesson for my real roster. Describe students by group and need, not by name.

Add supports for:
- English learners: [PROFICIENCY LEVELS], with language scaffolds (sentence frames, visuals, vocabulary pre-teach)
- Students with IEPs or 504s: [ACCOMMODATION TYPES], and how to meet these accommodations inside this lesson
- Students reading below grade level: scaffolds so the reading does not block the learning
- Early finishers: a meaningful extension that deepens the standard, not just "more work"

For each support, tell me which phase of the lesson it changes and what I do differently. Keep the core objective the same for everyone. Change the path, not the destination.
What good output looks like
  • Supports are tied to specific phases of the lesson, not bolted on at the end.
  • English-learner and IEP/504 supports change the path while keeping the same objective.
  • The extension deepens the standard instead of just adding more of the same work.

Verify before using AI's output

Step 5

Materials, Timing, and Exit Ticket

Turn the plan into something you can teach tomorrow: a materials list, a minute-by-minute flow, and an exit ticket that measures the objective.

Pull this together into a ready-to-teach package.

1. Materials list: everything I need to gather or print, including any handouts, slides, or manipulatives.
2. Minute-by-minute timing for the full period, including transitions, so the total matches my period length.
3. An exit ticket (2–4 quick items) that directly measures the objectives from Step 2 (not the hook, not the activity, the objective). Tell me what a "got it" answer looks like versus a "not yet" answer, so I can sort the stack fast.
What good output looks like
  • Materials list is specific enough to gather without rewriting anything.
  • Timing covers transitions and adds up to your real period length.
  • The exit ticket measures the stated objective and shows what “got it” versus “not yet” looks like.

Verify before using AI's output

Step 6

Match Your Good Examples, Avoid the Flop

Teach the AI what “good” and “bad” look like for you. Paste two lessons that worked and one that flopped, then have it revise toward your style.

What to substitute before pasting

  • [PASTE GOOD LESSON 1] A lesson that landed well with your students. Remove any student names or identifying details first.
  • [PASTE GOOD LESSON 2] A second lesson that worked. Same as before: strip names and identifiers.
  • [PASTE FLOPPED LESSON] A lesson that fell flat, so the AI can see what to avoid. Strip names and identifiers.
Here are examples of my own lessons so you can match my style.

Two that worked well:
[PASTE GOOD LESSON 1]
[PASTE GOOD LESSON 2]

One that flopped:
[PASTE FLOPPED LESSON]

Look at what makes the good ones work for me and my students, and what made the flop fall flat. Then revise the lesson plan you just built so it looks and feels like my strong lessons and avoids what sank the flop. Tell me the specific changes you made and why.
What good output looks like
  • The revision picks up specific moves from your strong lessons (your pacing, your structures, your style).
  • It names what made the flop fall flat and removes that pattern.
  • It lists the specific changes it made and why, instead of silently rewriting.

Verify before using AI's output

Step 7

Make AI Attack Its Own Plan

Have the AI find the holes before your students do. You own the final call on alignment; the AI only surfaces gaps.

Review this lesson plan for gaps. Be tough on it.

Find:
- Where the activities do not actually match the stated objectives (alignment slips)
- What the plan assumes my students already know that they may not
- Where the timing is unrealistic for my period
- Where a group (English learners, students with IEPs/504s, below-level readers) might get stuck or left out
- What check for understanding is missing or too weak to catch confusion
- What I need to decide as the teacher before I can teach this

Do not be polite. Find the holes. I will make the final call on alignment myself.
What good output looks like
  • Names specific alignment slips between activities and objectives, not vague concerns.
  • Flags assumptions about prior knowledge and spots where a group could get stuck.
  • Calls out the decisions only you, the teacher, can make.

Verify before using AI's output

Design or Redesign a Course with AI

A teacher needs to design a new course or rework one that is not landing the way they want. This is bigger than a single lesson. It covers the whole course or unit: what to teach, in what order, across multiple class sessions, with measurable objectives that line up with what the school or department requires. This exercise teaches a Description and Discernment method: you write a rich description of the course, then run a long collaborative conversation with AI where it proposes and you decide what to accept or reject. You stay the expert. AI is a thinking partner that never gets the final say.

Recommended tool

claude

Course design is iterative and runs across multiple sessions. It benefits from a long, collaborative conversation where AI proposes options and you discern which to keep, accepting, rejecting, and logging why as you go. Claude holds a long thread well and is strong at building on your description over many turns. If you are anchoring the design to an existing syllabus or textbook, consider NotebookLM so the conversation stays grounded in that specific source.

Bring to the exercise

  • The course or unit you are designing or redesigning
  • Your existing syllabus or materials, if you are revising rather than starting fresh
  • Your institutional requirements and the broader course goals your objectives must ladder up to
  • Your Educator Context Document (the "Teaching Context document") from Module 2
  • Your vision of what success looks like for this course
Step 1

Planning Your Approach (No AI Yet)

Get clear with yourself before you open any AI. The clearer your thinking now, the richer your description will be, and a rich description is what makes the rest of this work. AI cannot do this part for you.

Answer these for yourself before opening any AI:

1. Which course or unit am I designing or redesigning?
- Name it specifically. Is this a brand-new course, a full redesign, or one unit inside a larger course?

2. What work actually needs doing?
- Content selection: what topics and material go in?
- Sequencing: what order do they go in, across how many sessions?
- Objectives: what should students be able to do by the end, in measurable terms?
- Anything else this course needs?

3. Where is my expertise required, and where might AI help?
- Which decisions only I can make (pedagogy, what matters for these students, judgment calls)?
- Where could AI propose options, draft a first pass, or spot gaps I review?
- Be honest about the split. AI proposes; I discern.
What good output looks like
  • You can name the specific course or unit in one sentence.
  • You have listed the concrete work this design needs (content, sequencing, objectives).
  • You have a clear split between the decisions that are yours and the places AI can help.

Verify before using AI's output

Step 2

Initiating Collaboration

Open the conversation the right way. Hand AI your context, tell it what you are trying to do, and invite it to ask questions before it proposes anything. A good start sets up the whole back-and-forth.

What to substitute before pasting

  • [COURSE/UNIT] The specific course or unit you named in Stage 1.
  • [WHAT YOU'RE TRYING TO ACCOMPLISH] In plain terms: what do you want this course to do for students that it is not doing now?
  • [CONSTRAINTS] Time in the term, class length, resources, technology, and any institutional requirements that bound the design.
  • [FRESH OR REVISING] Say whether you are building from scratch or reworking material you already have.
I am a teacher designing a course, and I want to work with you as a thinking partner. You will propose ideas; I will decide what to keep. Ask me clarifying questions before you propose anything.

Here is my Teaching Context document. It describes how I teach, who my students are, and what I value:
[PASTE YOUR TEACHING CONTEXT DOCUMENT]

The course or unit I am working on:
[COURSE/UNIT]

What I am trying to accomplish:
[WHAT YOU'RE TRYING TO ACCOMPLISH]

My constraints (time in the term, class minutes, resources, technology, institutional requirements):
[CONSTRAINTS]

Am I starting fresh or revising existing material?
[FRESH OR REVISING; if revising, I will paste my current syllabus or materials next]

Before you suggest anything, ask me the questions you need answered to do this well.
What good output looks like
  • AI asks clarifying questions before it proposes a design.
  • AI references your Teaching Context document, not a generic course template.
  • AI shows it understands which course you mean and what you are trying to accomplish.

Verify before using AI's output

Step 3

Build Through Description & Discernment

This is the heart of the work. You make your description rich, then you discern: accept the good, redirect the weak, and never rubber-stamp. The richer you describe, the better AI proposes; the harder you discern, the more the course stays yours.

Now we build, in rounds. I will describe richly; you will propose; I will decide what to keep.

MAKE THE DESCRIPTION RICH
Help me think this through by responding to what I tell you:
- The specific challenges in this course: what tends to go wrong or fall flat.
- What makes this course unique: these students' needs, the prerequisites they bring (or don't), the outcomes that matter most.
- My vision of success: what it looks like when this course really works.
- My delegation thinking: what I want to collaborate with you on and why, and what I am keeping for myself.

TURN THE CONTENT MAP INTO OBJECTIVES
- Transform the content map into specific, measurable learning objectives.
- Check each objective against my institutional requirements.
- Make sure every objective ladders up to the broader course goals.
- Spot gaps and redundancies in the objectives and flag them for me.

PROPOSE AND LET ME DISCERN
- Suggest topics and concepts drawn from my context. Build on the good ideas; I will redirect the weak ones.
- For anything you propose, give me the reasoning so I can judge it. Do not just hand me a list.

LEARNING JOURNEY MAPPING
- Help me sequence 3-4 class sessions together, in order.
- At each transition point between sessions, take the students' perspective: "What might confuse students here? What scaffolding is needed?"
- Discuss how each session builds on the one before it.
What good output looks like
  • Objectives are specific and measurable, and each one ladders up to a broader course goal.
  • A sequence of 3-4 class sessions is mapped in order.
  • Likely confusion points are named at the transitions, each paired with the scaffolding it needs.

Verify before using AI's output

Step 4

Documentation & Reflection

Close the loop. Verify what AI gave you, write down the decisions and your reasoning, and be transparent about how AI was used. The discernment log is what proves the judgment was yours.

Help me document and reflect on this design.

VERIFY
- Walk through the design with me and check it for accuracy.
- Flag any claims, facts, or sources that need a citation or that I should confirm before I rely on them.

DOCUMENT MY DECISIONS
- Help me write down the key decisions I made and my reasoning for each.

THE DISCERNMENT LOG
- List the suggestions you made that I rejected, and help me note why I rejected each one. This is my discernment log: proof that I decided, not you.

AI-ROLE TRANSPARENCY STATEMENT
- Help me write a brief, honest statement about what role AI played in planning this course.
- Then help me think through how I might share this process with my students or colleagues.
What good output looks like
  • A written record of key decisions and the reasoning behind them.
  • A discernment log of the AI suggestions you rejected, with a reason for each.
  • A short, honest statement describing AI's role in the planning.

Verify before using AI's output

For Admins · PD, evaluation, operations, hard conversations

You're viewing the Teacher track. These Administrator exercises are also here. Switch to Administrator in the top bar to focus them.

Build a Professional Development Plan from an Identified Need

A school or district has identified a need. For example, classroom observations or survey data show that students are not talking enough during instruction. Teachers may be doing most of the talking. The administrator wants to increase student talk and decides teachers need support with cooperative learning, student discourse, or structured classroom discussion.

Recommended tool

claude

Claude is strong at multi-step operational planning, drafting structured progressions (theory → practice → coaching → impact), and producing the agenda + facilitator notes + walkthrough look-fors as a single coherent package.

Bring to the exercise

  • The identified need (and how it was identified: survey data, walkthroughs, assessment results, climate data, district priority)
  • The teacher group affected (grade level, department, experience level, strengths, challenges)
  • What teachers already know and what they need to learn
  • What the expected classroom practice should look like when PD is successful
  • What barriers may occur and likely teacher resistance
  • What follow-up support is available (coaching, observations, planning time)
  • What evidence will show progress
  • 10 examples of PD plans that worked + 1–2 that did not (redact any confidential info)
Step 1

Define the Need

Get AI to clarify the learning problem before designing a plan. AI cannot read your district context; you have to give it.

What to substitute before pasting

  • [INSERT NEED] One sentence: the learning problem you want PD to solve.
  • data source Pick the option that matches how you actually identified the need.
I am a school administrator designing a professional development plan. The identified need is: [INSERT NEED].

The need was identified through: [SURVEY DATA / CLASSROOM OBSERVATIONS / ASSESSMENT RESULTS / WALKTHROUGHS / CLIMATE DATA / DISTRICT PRIORITY / OTHER].

Help me clarify the learning problem. Identify what teachers likely need to know, what they need to be able to do, and what classroom practice should look like when the PD is successful. Ask me clarifying questions before drafting anything.
What good output looks like
  • Names what teachers need to know (knowledge), what they need to do (skills), and what students should experience (classroom practice).
  • Asks at least 2 clarifying questions before producing a plan.

Verify before using AI's output

Step 2

Add School and Teacher Context

Prevent generic PD by feeding AI the school/teacher context that shapes every decision.

What to substitute before pasting

  • [DISTRICT NAME] Your district’s name.
  • teacher concerns What teachers actually say in the lounge, in meetings, in 1:1s.
  • constraints Time, staffing, calendar, contract limits: anything that bounds the plan.
Here is my school context:
[GRADE LEVEL / SCHOOL TYPE / SIZE / STAFF MAKEUP / STUDENT POPULATION / CULTURE]

Here are the teachers involved:
[GRADE LEVEL / DEPARTMENT / EXPERIENCE LEVEL / STRENGTHS / CHALLENGES]

Here are common pain points or concerns these teachers may have:
[INSERT TEACHER CONCERNS]

Here are the [DISTRICT NAME] or building constraints:
[TIME / STAFFING / CALENDAR / TECHNOLOGY / EVALUATION REQUIREMENTS / UNION OR CONTRACT LIMITS]

Use this information to shape the PD plan so it fits our actual context. Do not produce generic PD.
What good output looks like
  • AI references your specific context (grade, size, contract limits) in its plan.
  • AI flags constraints that conflict with the proposed plan.

Verify before using AI's output

Step 3

Design the PD Progression

Move beyond a one-shot training day to a progression that ends with classroom impact.

Create a professional development progression that moves from theory to classroom practice.

Include:
- Initial training
- Practice or modeling
- Teacher planning time
- Classroom implementation
- Classroom visits or walkthroughs
- Coaching or feedback
- Follow-up training
- Evidence of implementation
- Evidence of impact

For each phase, name the purpose, the teacher action, the administrator action, and the artifact that proves the phase happened.
What good output looks like
  • Progression includes implementation, coaching, and evidence, not just training day.
  • Each phase has a named artifact (lesson plan, observation note, coaching memo).

Verify before using AI's output

Step 4

Build a Multi-Session Plan

Convert progression into actual sessions you can put on a calendar.

Turn this into a multi-session professional development plan.

For each session, include:
- Session title
- Purpose
- Learning outcomes (specific, observable)
- Activities (with rough timing)
- Materials needed
- Administrator role
- Teacher task
- Follow-up expectation
- Evidence that the session was successful

Sequence sessions so each builds on the last. Note prerequisites between sessions.
What good output looks like
  • 3–5 sessions with specific outcomes, not vague topics.
  • Each session names a follow-up expectation before the next session.

Verify before using AI's output

Step 5

Anticipate Resistance

Plan for the resistance you already know is coming. Resistance you don’t name will sink the plan.

Identify likely teacher resistance to this professional development plan.

Include resistance such as:
- "I already learned this."
- "This is just coming back around again."
- "This too shall pass."
- "You are taking away my autonomy."
- "I feel like I am being forced to teach like a robot."
- "I do not have time for this."
- "I am not doing this."

For each resistance point, give an administrator response that is:
- Direct (does not avoid the concern)
- Respectful (preserves teacher professionalism)
- Connected to the purpose of the PD (anchored in student needs)
- Specific (names the next step, not a pep talk)
What good output looks like
  • Each resistance pattern has a paired administrator response.
  • Responses anchor in student outcomes, not administrator authority.

Verify before using AI's output

Step 6

Create Buy-In

Adjust the plan to surface the WHY before the WHAT.

Revise the PD plan to increase teacher buy-in.

Include:
- How to explain the why
- How to connect the plan to student needs (use the original data we identified)
- How to preserve teacher professionalism (avoid making it feel like top-down compliance)
- How to communicate expectations clearly without scripting the classroom
- How to support hesitant teachers (extra coaching, peer pairing)
- How to hold resistant teachers accountable (clear expectations + documented follow-up)
What good output looks like
  • Buy-in plan separates hesitant from resistant, with different supports for each.
  • WHY is anchored in student-level data, not administrator preference.

Verify before using AI's output

Step 7

Create Administrator Tools

Generate the package you actually need to ship the PD.

Create the administrator tools needed to implement this professional development plan.

Include:
- Staff email announcing the PD
- PD agenda (with timing)
- Facilitator notes
- Teacher planning template
- Classroom implementation checklist
- Walkthrough look-fors (3–5 observable indicators)
- Coaching conversation guide
- Follow-up email after each session
- Evidence-of-impact plan (what data, who collects, when)
What good output looks like
  • Walkthrough look-fors are observable in 5 minutes, not vague.
  • Coaching guide names questions to ask, not statements to make.

Verify before using AI's output

Step 8

Review and Revise

Ask AI to attack its own plan before you ship it.

Review this PD plan for gaps. Be tough on it.

Identify:
- What is unclear
- What assumes too much (about teachers, students, or the building)
- What may create resistance you have not addressed
- What is missing from the follow-up
- What evidence should be collected that is not in the plan
- What I as the administrator need to decide before implementation

Do not be polite. Find the holes.
What good output looks like
  • AI lists 5+ specific gaps, not generic concerns.
  • AI flags decisions only the administrator can make.

Verify before using AI's output