AI Lesson Planning: The Complete Guide for Teachers

AI lesson planning is the practice of using an AI tool to draft, adapt and extend lesson plans from information a teacher provides: the class, the subject, the learning objective, the time available and any standards or constraints. Done well, it shortens the slowest part of planning (getting from a blank page to a workable first draft) while leaving the decisions that matter with the teacher.

This guide is the long version. It covers what AI is good and bad at in planning, a repeatable workflow you can use for a single lesson or a whole unit, how to check quality, how to differentiate, and how to connect the plan to the assessment that follows it. It is written for classroom teachers, but heads of department and instructional coaches will find the checklists useful for supporting colleagues.

The short answer

If you only read one section, read this:

1. Start from the objective, not the activity. Tell the AI exactly what students should know or be able to do by the end of the lesson.
2. Give real context. Grade level, subject, lesson length, prior knowledge, standards and anything specific about your class.
3. Treat the output as a draft. Check accuracy, sequencing, timing and fit with your curriculum before you use it.
4. Plan the check for understanding at the same time. An exit ticket or short quiz, written alongside the plan, tells you whether the lesson worked.
5. Keep everything connected. The plan, the resources, the quiz and the feedback should come from the same context so tomorrow’s lesson can respond to today’s evidence.

The rest of this guide explains each step in detail.

What AI is good at in lesson planning, and what it is not

It helps to be clear-eyed about where AI saves time and where it creates risk.

Where AI helps

• First drafts. Turning an objective into a sequenced outline with an opener, explanation, guided practice, independent practice and a plenary takes seconds rather than half an hour.
• Variations. Producing the same activity at three levels of challenge, or the same explanation with a different worked example, is tedious by hand and quick with AI.
• Question writing. Drafting hinge questions, retrieval questions, discussion prompts and exit tickets aligned to a specific objective.
• Adapting existing material. Turning a reading, a video or your own slides into a lesson outline, a glossary or a set of comprehension questions.
• Admin-shaped tasks. Cover lesson notes, parent-friendly summaries, rubric drafts and report comment starting points.

Where teachers must stay in charge

• Knowing the class. AI does not know that period 4 struggles after lunch, which students need a seating change, or what happened in last week’s lesson.
• Curriculum fit. Your scheme of work, your exam board or district standards and your department’s agreed sequence are decisions, not prompts.
• Accuracy. AI can state things confidently that are wrong, out of date or subtly misleading, especially in specialised subjects.
• Pedagogical judgement. Choosing when to model, when to let students struggle, and when to stop and reteach is the core of teaching.

Guidance from bodies such as UNESCO frames AI in education the same way: a tool that should be used in a human-centred way, with teachers accountable for what reaches students. A good AI lesson-planning workflow is designed around that principle.

The complete AI lesson-planning workflow

The workflow below works for a single lesson. Later sections extend it to units, differentiation and cover lessons.

Step 1: Write a precise learning objective

The single biggest improvement most teachers can make to AI lesson plans is a sharper objective. Compare:

• Weak: “Photosynthesis.”
• Better: “Students can explain how light intensity affects the rate of photosynthesis and interpret a simple rate graph.”

The second version tells the AI what knowledge matters, what students must do with it, and what evidence of success looks like. It also gives you something concrete to check the plan against.

A useful pattern is: students can [verb] [content] [condition or context]. Verbs such as explain, compare, calculate, justify and evaluate produce more useful plans than understand or know, because they imply a task.

This is essentially backward design, popularised by Wiggins and McTighe’s Understanding by Design: decide what students should be able to do, decide how you will know, and only then plan the activities. AI makes the third step faster; it does not remove the need for the first two.

Step 2: Give the AI the context a colleague would need

Imagine handing the lesson to an experienced colleague who has never met your class. What would they need to know? That is what the AI needs too.

• Grade level and subject.
• Lesson length. A 50-minute and a 90-minute lesson need different structures.
• Prior knowledge. What did students learn last lesson? What misconceptions came up?
• Standards or specification points you need to cover.
• Class profile. Mixed attainment, multilingual learners, students with support plans, a class that needs short tasks, and so on, without identifying individual students.
• Constraints. No devices, a practical lab, a room without a projector, a mock exam next week.
• Format preferences. The headings your school uses, or a specific lesson structure.

In Duetoday’s Lesson Plan Generator, the core fields are grade level, subject, topic and duration, with optional fields for standards and additional criteria. The additional criteria box is where the class profile and constraints belong. The more specific you are there, the less editing you do later.

Privacy note: describe needs, not names. “Three students are working below grade level in reading” is useful context; a student’s name, diagnosis or personal details are not needed and should not be pasted into any AI tool unless your school’s policy explicitly allows it.

Step 3: Generate a first draft and read it like a reviewer

Once you generate the plan, read it as if you were observing a colleague’s lesson. Ask:

• Does every phase of the lesson serve the objective?
• Is the explanation accurate and pitched at the right level?
• Is there enough modelling before independent work?
• Are there planned checks for understanding, not just at the end?
• Does the timing add up to the lesson length, with time for transitions?
• Is anything there that you would never do with this class?

It is normal to change a quarter or more of an AI draft. The time saving comes from editing rather than writing from scratch, not from using the draft unchanged.

Step 4: Refine through conversation, not by starting over

Most AI planners, including Duetoday’s, let you keep refining the plan rather than regenerating it. Targeted follow-ups work better than vague ones:

• “Shorten the starter to five minutes and make it a retrieval quiz on last lesson’s content.”
• “Add a worked example before the independent task, using a different context from the one in the explanation.”
• “Replace the group activity with paired work; this class works better in pairs.”
• “Add two hinge questions after the explanation with plausible wrong answers that reveal the common misconception about X.”

Each instruction changes one thing, so you stay in control of the result.

Step 5: Build the resources from the same context

A lesson plan on its own rarely solves the workload problem. You still need the slides, the worksheet, the questions, the quiz and possibly a rubric. This is where many AI workflows fall apart: teachers generate a plan in one tool, then rebuild the context from scratch in three others.

Keeping the plan and the resources together saves that re-entry. In Duetoday, the teacher area brings the Lesson Plan Generator together with the Assessment Builder, Rubric Generator, Report Card Comments, flashcards and live classroom quizzes, and Studio for generating further resources. Generating from the same objective and content means the worksheet, quiz and rubric line up with what you actually taught.

Step 6: Plan the check for understanding before you teach

Decide now how you will know whether the lesson worked. Options include:

• An exit ticket of two or three questions tied directly to the objective.
• Hinge questions at the midpoint, so you can decide whether to move on.
• A short live quiz at the start of the next lesson to check retention.
• A mini-whiteboard routine where every student answers at once.

Writing these at planning time keeps them aligned with the objective and stops the plenary becoming an afterthought. For more on this, see the companion guide to AI for formative assessment.

Step 7: Teach, then close the loop

After the lesson, spend two minutes noting what happened. Which questions did most students miss? Which part ran over? Feed that back into the next plan as prior knowledge and misconceptions. This is the step that turns AI from a draft generator into a genuinely responsive planning system, and it is the one most often skipped.

A worked example

Here is how the workflow looks for a single lesson. The inputs below are the kind of thing you might type into a lesson planner.

Inputs

• Grade level: Year 9 (age 13–14)
• Subject: Science (biology)
• Topic: Factors affecting the rate of photosynthesis
• Duration: 60 minutes
• Standards: your specification point on photosynthesis rate
• Additional criteria: mixed-attainment class; several students find graph interpretation hard; practical lab available but limited equipment; last lesson covered the word equation for photosynthesis and some students confused it with respiration.

What to look for in the draft

• A starter that retrieves the word equation and explicitly separates it from respiration.
• A clear explanation of limiting factors with one modelled graph before students interpret one alone.
• A short practical or demonstration that fits the equipment.
• At least one hinge question after the explanation.
• Independent graph interpretation at more than one level of support.
• An exit ticket with one graph question and one explanation question.

Edits a teacher might make

• Swap the suggested practical for a demonstration because of equipment limits.
• Replace a generic graph with one from the class’s textbook so the format matches assessments.
• Add sentence starters for the explanation question.
• Cut a discussion activity that would push the lesson over time.

Resources built from the same context

• A worksheet with three graphs at increasing difficulty.
• A five-question live quiz for the start of next lesson.
• A short rubric for the extended explanation question.

None of these steps require clever prompting. They require clear inputs, a critical read and a connected set of tools.

Prompts and inputs that work

If you are using a free-text AI tool rather than a structured planner, these prompt patterns produce more useful drafts. In a structured planner, the same information goes into the fields and the additional criteria box.

Full lesson

Plan a [length]-minute [subject] lesson for [grade level]. Objective: students can [verb + content]. Prior knowledge: [what they already know]. Common misconception: [misconception]. Class context: [profile and constraints, no names]. Include a retrieval starter, an explanation with one worked example, guided practice, independent practice at two levels of challenge, two hinge questions and a three-question exit ticket. Give timings for each phase.

Improving a plan you already have

Here is my lesson plan. Keep the structure and objective. Suggest three changes that would increase student thinking during the explanation phase, and explain why each would help.

Questions only

Write five multiple-choice hinge questions on [objective] for [grade level]. Each wrong answer should reflect a specific misconception. List the misconception next to each distractor.

Cover lesson

Turn this lesson plan into cover instructions a non-specialist could follow. Use plain language, include timings, list the resources needed and say exactly what students should hand in.

Each of these keeps the objective and context explicit, which is what drives quality.

Quality checklist before you teach

Use this list on every AI-drafted plan. It takes three to five minutes and catches the problems that matter.

1. Objective alignment. Every activity serves the objective. Anything that does not is cut.
2. Accuracy. Facts, definitions, formulas, dates and examples are correct and match the terminology your course uses.
3. Sequencing. New content is introduced in small steps, with modelling before independent practice.
4. Checks for understanding. There is at least one point mid-lesson where you will know whether to move on.
5. Timing. Phases add up to the lesson length, with realistic transitions.
6. Accessibility. Reading load, vocabulary and task instructions are appropriate for the whole class.
7. Resources. Everything the plan refers to exists, or you know how you will create it.
8. Fit with your class. You can picture this lesson working with these students in this room.

Rosenshine’s Principles of Instruction are a useful lens here: daily review, small steps, lots of questions, modelling, guided practice and checking understanding. If an AI plan is missing several of these, ask for them explicitly.

Differentiation with AI

Differentiation is one of the areas where AI saves the most time, because producing the same task at several levels by hand is slow.

Practical approaches

• Same objective, different support. Ask for one core task with a scaffolded version (sentence starters, worked examples, partially completed tables) and an extension version (an extra condition, a “what if” question, a justification).
• Vocabulary support. Generate a short glossary with student-friendly definitions and an example sentence for each key term.
• Reading level adjustments. Rewrite a source text at a more accessible level while keeping the key content and terminology.
• Multilingual learners. Ask for visual supports, key vocabulary with simple definitions and sentence frames for discussion and writing.

Guardrails

• Keep the objective the same for everyone unless you have a deliberate reason not to. Differentiation should change the route, not lower the destination.
• Check that simplified texts have not lost accuracy.
• Avoid labelling groups in ways students will see; describe tasks by colour or number, not ability.

For more detailed strategies, see AI differentiation strategies for teachers.

Planning units, not just lessons

Lesson-by-lesson planning with AI is useful, but the bigger gain comes from planning a sequence.

1. Start with the end-of-unit assessment. What should students be able to do after the unit? Draft or select the assessment first.
2. Break the unit into lesson objectives. Ask AI to propose a sequence, then reorder it to match your curriculum and the way concepts build.
3. Identify retrieval points. Plan where earlier content will be revisited in later lessons, rather than leaving it until revision.
4. Generate lessons in order, using each lesson’s evidence (what students got wrong) as prior knowledge for the next.
5. Build a shared bank of questions for the unit that you can reuse in starters, quizzes and revision.

This turns AI from a one-off draft generator into a planning assistant for the whole term.

Connecting the plan to assessment and feedback

A lesson plan is only as good as the evidence it produces. The most effective AI planning workflows are built around a loop:

1. Plan the lesson and its checks for understanding together.
2. Teach and collect evidence through hinge questions, exit tickets or a live quiz.
3. Analyse the responses: which questions did most students miss, and why?
4. Respond in the next lesson with targeted reteaching or practice.

AI can help at every stage. Live classroom quizzes give you whole-class evidence in minutes; see the guide to live classroom quizzes for how to run them well. A rubric drafted from the same objective keeps marking consistent. Report comment tools can turn your notes on a student’s progress into a first draft you then personalise.

The common thread is that the plan, the assessment and the feedback share the same objective and content. When they come from different tools with different inputs, they drift apart.

Cover lessons and substitute plans

Cover lessons are a natural fit for AI, because the main challenge is clarity, not creativity.

• Write for a non-specialist. Ask for instructions that someone outside your subject could follow.
• Make every task self-contained. Students should be able to complete the work without new teaching.
• State what success looks like and what should be handed in.
• Keep it retrieval-heavy. Practice on content students already know works better than new content without a specialist.

For the wider admin around cover, see AI teacher workflow automation.

Common mistakes to avoid

Vague inputs. “Make a lesson on fractions” produces a generic lesson. Specific objectives and context produce specific lessons.

Using the draft unchanged. AI drafts are first drafts. Teaching one without reading it carefully is the fastest way to lose trust in the tool and, more importantly, to teach something inaccurate.

Overloaded lessons. AI tends to include too much. If a plan has six activities in fifty minutes, cut it down to the ones that matter most.

Activities without purpose. A creative activity that does not serve the objective is a distraction. Ask “what will students be thinking about during this?” for every phase.

Forgetting the evidence. A plan without checks for understanding gives you no information about whether students learned anything.

Pasting personal data. Student names, assessment records and support plans should not go into AI tools unless your school has approved that use.

Planning in isolation. If the plan, the worksheet and the quiz come from three disconnected tools, you spend the time you saved on re-entering context.

Using AI responsibly in your school

Most schools now have, or are writing, a policy on AI. A responsible AI lesson-planning practice usually includes:

• Following your school’s policy on which tools are approved and what data can be shared.
• Keeping teachers accountable for everything that reaches students.
• Being transparent with colleagues and, where appropriate, students and parents about how AI is used.
• Checking for bias in examples, names and contexts, and choosing materials that reflect your students.
• Building AI literacy. Students also need to learn how AI works and where it goes wrong; the AI literacy guide covers this.

How to plan lessons in Duetoday

If you want to try the workflow in Duetoday, here is the practical version.

1. Open the teacher area on a laptop or desktop and choose the Lesson Plan Generator.
2. Fill in the core fields: grade level, subject, topic and duration.
3. Add standards you need to align to, and use additional criteria for prior knowledge, misconceptions, class profile and constraints.
4. Generate the plan and read it against the quality checklist above.
5. Refine it with specific follow-up instructions rather than regenerating from scratch.
6. Build the resources from the same context: an assessment with the Assessment Builder, a marking guide with the Rubric Generator, or flashcards and further materials in Studio.
7. Create a live quiz to check understanding at the end of the lesson or the start of the next one. Students join at duetoday.ai/join with a code; see how to run a live classroom quiz.
8. Share resources with your class through a study room, so students have the materials after the lesson.

Duetoday has no free plan; every account starts with a 7-day free trial (card required, cancel before day 7 to pay nothing), then $15 a month or $99 a year. See pricing and the teachers page for details.

Frequently asked questions

Is it acceptable for teachers to use AI to plan lessons?

In most schools, yes, provided you follow your school’s AI policy and remain responsible for what you teach. Using AI to draft a plan is similar to adapting a plan from a colleague or a published scheme: the professional judgement is in the choices you make about it.

Will AI lesson plans be aligned to my standards?

Only if you tell the AI which standards to align to and check the result. A planner with a standards field makes this easier, but you should still confirm that the objectives and activities genuinely address the standard rather than mentioning it.

How much time does AI lesson planning save?

It depends on how you plan now and how much editing the drafts need. The biggest savings usually come from producing the resources around the plan (questions, differentiated tasks, quizzes and rubrics), not from the plan itself.

Can AI plan a whole unit?

It can propose a sequence of lessons and objectives, which is a useful starting point. You should reorder and adjust it to match your curriculum, and plan lessons one at a time so each one can respond to evidence from the last.

What should I never put into an AI lesson planner?

Student names, personal details, assessment records or support plans, unless your school has explicitly approved that use. Describe needs in general terms instead.

Does AI lesson planning work for every subject?

The workflow works across subjects, but checking accuracy matters more in specialised areas such as advanced sciences, mathematics and languages, where small errors are easy to miss. In practical and creative subjects, expect to edit activities more heavily to fit your equipment and space.

Further reading

• AI for formative assessment: a practical teacher guide
• Live classroom quizzes: a complete guide for teachers
• AI differentiation strategies for teachers
• Best AI for lesson planning: tools compared
• AI teacher workflow automation: a practical guide
• Duetoday for teachers