How to Build a Curriculum-Aligned Lesson Plan With AI

For a new lesson, start with the planning brief. To improve a lesson you already teach, use the revision prompt. The science, mathematics, and reading examples show how the tasks and assessment change by subject. You can use a school-approved AI writing tool; students do not need AI accounts.

1. Start with the curriculum your class actually follows

Gather the current outcome from your school, district, exam board, or curriculum authority. Add the relevant unit sequence and any assessment guidance. A national standard describes an expectation; your local curriculum determines where this lesson belongs and what students have already encountered.

The worked examples use US standards. For another curriculum, supply its actual outcome, year group, assessment guidance, and place in your school's sequence. Compare the required content and student action before borrowing a task. Similar topic names or age groups do not establish that two curricula expect the same learning.

Give AI the wording and the source location. Asking it to recall a code invites a plan built around an unchecked interpretation. If you use a PDF, copy the relevant passage and its page number. Check that tables, footnotes, and mathematical notation survived extraction.

If the lesson also follows a textbook chapter, include the relevant pages as teaching material and check them against the adopted outcome. Give the model the selected extract and its adjacent guidance so it can see the conditions attached to the outcome.

If your curriculum gives a broad outcome such as “evaluate historical sources,” choose a manageable part for this lesson: compare two supplied accounts, identify a disagreement, and justify which details support a particular interpretation. Record that scope as your lesson decision. Do not attribute your narrower wording to the curriculum authority.

Read the curriculum extract below. Do not draft activities yet.

Return:
1. The student action and the content it applies to.
2. Conditions, assessment limits, and necessary prior knowledge.
3. A manageable objective for one lesson.
4. Student work that would demonstrate that objective.
5. Unanswered questions that could change the plan.

Separate source requirements from your teaching suggestions.
Use the supplied source and page or section for each requirement.
Do not invent curriculum codes, quotations, or local policy.

Curriculum extract: [paste checked wording]
Source and location: [URL, document edition, page or section]
Place in unit: [previous learning and next lesson]

2. Decide what you will assess before choosing activities

Write an exit task that students can complete individually. Then decide what a satisfactory response contains. This gives you something concrete against which to judge an AI-generated lesson.

“Understand fractions” leaves too many decisions unresolved. “Solve a word problem involving two fractions of the same whole with the same denominator, and show the relationship with a model and equation” names observable work. Your success criteria should describe mathematical reasoning, not neat colouring or completing every question.

Separate the objective from the activity. A card sort is a teaching format. The learning evidence might be a student's written justification for placing one card in a category. Group discussion can prepare that response, but a shared poster gives limited information about each learner.

Keep the assessment small enough to inspect. For an introductory lesson, one well-chosen individual task may be more informative than a long worksheet students cannot finish. A single lesson usually provides evidence about part of a larger standard; record the remaining coverage in your unit plan.

For the science example below, a possible sequence is to introduce characteristic properties in an earlier lesson, compare supplied records in this lesson, and later check transfer using unfamiliar data. If the prerequisite lesson showed that students still confuse a property's value with the amount of material, revisit that distinction before asking them to judge a reaction.

3. Give AI a brief that includes the classroom constraints

Supply usable teaching time, class size, resources, prerequisite knowledge, and relevant access needs. Include the time for distributing materials and collecting work. State whether students can work online and whether the lesson must function entirely on paper.

Describe learning needs without uploading names, diagnoses, individual education plans, or identifiable work to an unapproved service. “Several learners need vocabulary rehearsal; one response route must work without handwriting” is enough for a draft. Apply individual arrangements through your school's approved process.

Draft a lesson using this brief. Use only the supplied curriculum
and teaching materials for factual content. List missing essentials
that need a teacher's decision and explain what depends on them.

Subject and class: [ ]
Adopted outcome, exact extract, and source location: [ ]
Lesson objective: [ ]
Prior learning and a likely misconception: [ ]
Total minutes, including transitions and collection: [ ]
Class size and grouping: [ ]
Available texts, data, equipment, and printing: [ ]
Language and access supports, without identifying learners: [ ]
Individual exit task: [ ]
Success criteria and answer key: [ ]
Restrictions or required school procedures: [ ]

Return:
- A table linking each objective, task, and student response.
- A timed sequence with teacher moves and student actions.
- The exact student questions and checked model answers.
- A question to check prior knowledge and how to respond to mistakes.
- Supported practice and a more demanding extension.
- An independent exit task with scoring guidance.
- A shorter version and a paper-only fallback.

Make the minutes add to the supplied total. Label proposed timings
as estimates. Do not claim classroom testing or school approval.

If the outcome or source material is missing, fill that gap before accepting a finished plan. Smaller uncertainties, such as which approved response tool students will use, can remain visible in a draft. You do not need to stop all planning for every unresolved detail.

Ask AI for the alignment table first when the brief is complex. Correcting a mismatch there is easier than rebuilding a long lesson with worksheets and slides.

If you already have a lesson plan

Paste the existing plan beside the outcome and answer key. Ask AI to identify which parts already teach the required learning, which parts need revision, and what evidence is missing. Preserve activities that work for your class. A new worksheet or a complete rewrite is unnecessary when changing the question or adding an individual response will solve the mismatch.

For example, if a draft assesses a group poster that lists reaction signs, ask for a revision that requires each student to compare the supplied property records and explain a reaction inference. Keep useful preparation for that task and replace the assessment that only checks recall.

Compare this existing lesson with the supplied curriculum outcome.
Identify the parts to keep and the specific mismatches to fix.
For each mismatch, explain what student thinking is missing.
Revise the affected questions, teaching steps, and answer key.
Keep the stated class time, resources, and response supports.

Existing lesson and answer key: [paste]
Curriculum extract and source: [paste]
Class constraints: [paste]

A filled-in brief for the science example

The input below puts the method into one prompt. Its curriculum summary has been checked against the NGSS sources linked in the next section; it is a working paraphrase. The lesson and practice records are illustrative. You can compare a proposed response with the task map and complete lesson below.

Plan a 52-minute Grade 7 science lesson for 28 students in seven groups.
Use printed records, pencils, and a board. Include distribution,
transitions, and collection. Students will not handle chemicals.

Curriculum basis: NGSS MS-PS1-2 and its official evidence statement.
Working summary: organise before-and-after characteristic-property
data, identify patterns, infer whether a reaction occurred, and
relate changed substances to rearrangement of atoms.
This lesson addresses part of the standard within a larger unit.

Prior learning: pure substances, characteristic properties,
units, and the idea of atomic rearrangement.
Objective: compare the records and explain a reaction inference
using property evidence and atomic reasoning.

Constructed practice records, not actual chemical measurements:
Case 1, before A: density 1.20 g/cm³; melting point 60°C; soluble.
Case 1, before B: density 2.00 g/cm³; melting point 105°C; soluble.
Case 1, isolated after-sample C: 1.60 g/cm³; 160°C; insoluble.
Case 2, D before and after recovery: 1.50 g/cm³; 80°C; soluble.
Case 3: bubbles observed; no characteristic-property records.
For Cases 1 and 2, assume pure, separately isolated samples,
density at 20°C, melting points at the same pressure,
and solubility in water at 20°C.

Fresh exit card: E melts at 45°C; F at 90°C; isolated after-sample
G at 140°C. Other property records and purity checks are missing.
Ask what this supports, how atoms relate to the inference,
what limits the conclusion, and which comparison would help.

Support: clear row labels, vocabulary rehearsal, and an appropriate
oral response option that retains the student's own reasoning.
Assessment: individual explanations using the supplied records.
Check comparisons, reaction inference, atomic reasoning, and limits.

First return a table connecting the objective, student tasks,
individual responses, and what each response will show.
Flag missing information. Do not invent additional measurements.

Check that the proposed tasks require students to interpret the records. If the model substitutes a demonstration, vocabulary quiz, or decorative poster, ask it to replace that part and name the individual response you need. Then request the teaching sequence:

Using the reviewed task map, write the complete 52-minute lesson.
Include a prior-knowledge question, teacher modelling, supported
practice, a check during the lesson, feedback, and an individual
exit task. Give exact student questions and model answers.
Add a rubric scoring each of the four assessment areas from 0 to 2.
Keep the supplied records and comparison conditions unchanged.
Show the time for each part, including transitions and collection.
If the plan is too long, remove an activity and retain time for
students to explain their reasoning and use feedback.

When revising the result, name the section to change and restate the constraints it must preserve. For instance, ask to shorten the card sort while keeping the individual explanation and collection time. Recheck the affected answers and the total minutes after the change.

4. Worked science example: interpret evidence of a chemical reaction

For a middle-school science example, NGSS MS-PS1-2 concerns interpreting substance-property data before and after interaction to decide whether a reaction occurred. Its assessment boundary lists density, melting point, boiling point, solubility, flammability, and odor. A table of bubbles, mass, and temperature alone does not cover that stated property scope.

The official evidence statement, page 2, also asks students to relate changed properties to atoms rearranging into new substances. Keep that age-appropriate reasoning in the task. Detailed bond calculations are unnecessary for the example below, but removing all atomic reasoning would omit part of the evidence expectation.

This is one formative lesson within a unit, not a complete assessment of every listed property or the full standard. It uses supplied records so students can spend the period organising evidence and explaining an inference. Listing flammability and odor in a standard does not require students to perform flame or smell tests.

The lesson brief

Plan for a fictional class of 28 students working in seven groups, with individual exit responses. Allow 52 minutes. Students have already encountered pure substances, characteristic properties, and the idea that a chemical reaction rearranges atoms. The teacher needs printed property cards, blank comparison tables, pencils, and a board. No chemical apparatus is needed.

Use this lesson objective: “I can organise before-and-after property records, decide whether they support the formation of a new substance, and explain my decision using the data and the idea of atomic rearrangement.” Treat evaluating missing evidence as an additional reasoning goal.

Practice data and the conditions that make a comparison meaningful

The values below are deliberately constructed practice data. A, B, C, and D are placeholders, not identified chemicals. These numbers are not experimental measurements or reference values. For the exercise, stipulate that each record concerns a pure, separately isolated sample; density is compared at the same temperature, melting points at the same pressure, and solubility using the same solvent and temperature. These assumptions belong on the student sheet.

Give students Case 3 as a separate card: “Bubbles appeared during an interaction. No characteristic-property records were collected.” Its purpose is to make students recognise insufficient evidence. The card cannot support the same analysis as Cases 1 and 2.

Case 1 is constructed to support the formation of a new substance under the stated assumptions. Case 2 provides no observed property difference supporting that conclusion. It does not prove that every possible chemical change has been ruled out. Before using real records, replace the practice cards with checked curriculum data and preserve their source, sample descriptions, and measurement conditions.

Map the objective to student work

For this lesson, ask for two property comparisons to make the evidence visible. That number is a teaching choice, not a quoted NGSS requirement. A copied statement about atoms without any reference to the supplied data is also incomplete.

A complete 52-minute sequence

The durations add to 52 minutes and include distribution and collection. They remain estimates. Rehearse the explanation and check the reading load before teaching; shorten the number of cards if necessary.

Student questions and example answers

Ask students to complete these questions after organising the records:

1. Which properties of C differ from A and B? Identify the sample and value for each comparison.
2. Does Case 1 support formation of a new substance? Explain how the comparisons support your decision.
3. How does rearrangement of atoms help explain the formation of substances with different properties?
4. What can you conclude from Case 2? What would be too strong a conclusion?
5. What is missing from Case 3? Propose one relevant before-and-after property comparison.

A strong illustrative Case 1 response is: “The records support a reaction. C melts at 160°C, compared with 60°C for A and 105°C for B. C is also insoluble in water under the stated conditions, while both starting substances are soluble. These differences support formation of a new substance. In a chemical reaction, atoms from the starting substances rearrange to form new substances with different properties. This interpretation depends on C being a pure, isolated sample and on the comparisons using the stated conditions.”

A partial response is: “C has a higher melting point, so there was a reaction.” It identifies relevant data, but should compare C with both starting substances and explain why a new property pattern matters. Prompt the student to use the solubility record and connect the inference to atomic rearrangement.

An unsupported response is: “There was a reaction because bubbles always mean a new substance.” Ask the student which characteristic-property record supports that statement. If they cannot identify one, return to Case 3 and practise describing the evidence as insufficient.

For Case 2, accept: “The recorded properties match before and after recovery, so these records do not support a new substance. They cannot establish that every possible reaction has been excluded.” For Case 3, accept a proposal to compare a suitable characteristic property of identified starting and resulting samples under consistent conditions. Students should not invent measurements to complete the card.

An eight-point rubric and an independent exit task

This rubric is for feedback during learning. Its final row adds an emphasis on evidence limits; the score is not an official NGSS proficiency level. Record which parts need work because students with the same total may need different follow-up teaching.

For the exit task, give students a fresh card: “E melts at 45°C and F at 90°C. An isolated after-interaction sample G melts at 140°C. No other property records or purity checks are supplied.” Ask for the inference this supports, an atomic explanation, and a reason to seek another comparison. The intended answer is that the difference supports possible formation of a new substance, while the limited records and unknown purity prevent a conclusive interpretation. Ask students which additional characteristic-property record would help. Use this shorter task to check the inference, atomic reasoning, and limits; use the earlier table for the two-property comparison criterion.

If students read values accurately but omit the atomic connection, revisit the relationship between changed substances and rearranged atoms. If they cannot match corresponding rows, reteach table reading first. If most students treat Case 3 as conclusive, use another incomplete record before introducing a new phenomenon.

5. Adapt the method for a mathematics lesson

Common Core 4.NF.B.3.d addresses fraction addition and subtraction word problems with like denominators and a common whole. Keep those conditions in the AI brief. General addition of fractions with unlike denominators is not a requirement of this Grade 4 domain.

For a proposed 40-minute lesson, assume students can identify equal parts and interpret a fraction of one whole. Check that prerequisite with a strip divided into eight equal parts: ask what one part represents and why unequal pieces would make the representation misleading.

Use this practice problem: “One garden bed has eight equal sections. Three eighths are planted with beans and another two eighths with carrots. What fraction of the bed is planted with these crops? What fraction remains?” Students represent one bed, write 3/8 + 2/8 = 5/8, and then 8/8 − 5/8 = 3/8. The model should show that the parts remain eighths.

Allow 5 minutes for the prerequisite check, 8 for modelling, 10 for paired practice, 7 for discussing an error, 7 for an independent task, and 3 for collection and the next-step decision. During the error discussion, show the illustrative answer 3/8 + 2/8 = 5/16. Ask students to identify what the denominator counts and to test the answer against the garden model.

For the independent task, use: “A ribbon is divided into eight equal lengths. Lina uses five eighths. What fraction remains? Draw or describe a representation and write an equation.” The answer is 8/8 − 5/8 = 3/8. Check the operation, the common whole, and the representation separately.

Offer a pre-drawn eight-part strip to learners who need help organising the model. They still decide which parts to mark and which operation fits the story. For an extension, ask students to create two different garden stories whose planted total is 7/8 and explain both equations. Increasing the reasoning demand keeps the work within the selected content.

Review these fraction tasks against the supplied 4.NF.B.3.d extract.
Check the common whole, like denominators, story meaning, model,
equation, and answer. Solve every task independently.
Identify any prerequisite being assumed or harder content being added.
For each error, give the original task, the mismatch, and a correction.
Do not accept a correct numerical answer as sufficient evidence
that the story and representation agree.

Tasks and proposed answers: [paste]

6. Adapt the method for a reading lesson

Common Core RI.6.1 requires textual evidence for analysis of explicit meaning and inferences. Plan questions that make students explain how the selected evidence supports a claim. A quotation copied beside an opinion gives you little evidence of that connection.

Choose an approved informational text and give AI the complete relevant passage with paragraph or line numbers. Verify every proposed answer against that version. If the model has only a title or summary, it cannot reliably provide exact passage evidence.

The short original instructional passage below is for demonstrating the question design. It is not a complete Grade 6 reading selection or a substitute for checking text complexity.

[1] A teacher begins by deciding what a lesson should help students do. She writes a short individual task before choosing the classroom activities.

[2] During practice, she reads students' explanations. A correct answer with no explanation may leave her uncertain about how the student reached it.

[3] She uses the work to choose the next example. Students then revise an explanation or try a new question, so they have an opportunity to use the feedback.

Ask an explicit question: “What does the teacher write before choosing the activities?” The answer is the short individual task in paragraph 1. Then ask an inference question: “Why might collecting correct answers alone be insufficient for her next teaching decision?”

A supported inference is: “She needs information about students' reasoning to choose useful follow-up work. Paragraph 2 says a correct answer may leave her uncertain about how it was reached, and paragraph 3 says she uses the work to choose the next example.” A weaker response, “She wants everyone to get full marks,” adds a motive the passage does not establish.

For a proposed 45-minute lesson using a suitably challenging approved text, allocate 6 minutes to reading and essential vocabulary, 8 to modelling evidence selection, 10 to paired analysis, 6 to comparing a supported and unsupported inference, 10 to an individual response, and 5 to feedback and collection. Model one question and reserve a different inference for the individual task.

Score whether the inference is defensible, the selected evidence is relevant and accurate, and the explanation connects them. Require both explicit analysis and inference across the lesson, but do not invent a rule that the standard always requires exactly two quotations. A targeted prompt can specify two details as a local teaching choice.

For an extension, give students two plausible interpretations and ask which is better supported, or ask what the passage leaves unresolved. If students select irrelevant sentences, model how each candidate detail relates to the claim. If the evidence is relevant but the explanation is missing, ask them to complete “This detail supports my inference because...” without supplying the reasoning.

Using only the numbered passage below, draft one explicit question
and two inference questions. For each, give a model response,
its exact evidence locations, and an explanation of the connection.

Quote only wording present in the passage. Mark an inference
unsupported if the text cannot justify it. Include one plausible
wrong response and a follow-up question that helps a learner
inspect the evidence without revealing the answer.

Passage: [paste approved text with paragraph or line numbers]

7. Adjust access and challenge while keeping the learning visible

CAST's Universal Design for Learning Guidelines include options for representing information and expressing learning. Use those options to address the barrier in a particular task. Writing “differentiate as needed” at the end of a lesson does not tell a teacher what to change.

Choose supports according to the skill the task is meant to assess. Reading a science data table aloud can help reveal a student's science reasoning when decoding is incidental. If the lesson assesses independent reading or a specified writing skill, a changed response route may alter what the task measures. Follow the relevant assessment arrangements and individual plans.

For a class spanning different attainment levels, you can use a shared phenomenon or text with different prompts. State what each group is expected to learn and how you will assess it if the goals differ. Do not label a reduced objective “the same standard” merely because everyone uses the same worksheet.

8. Handle a short period, limited equipment, or absent students

For a 30-minute version of the science lesson, use 4 minutes for the diagnostic question, 5 for vocabulary and modelling, 8 for Case 1 analysis, 5 for feedback and revision, 6 for an independent explanation, and 2 for collection. Move the full Case 2 and Case 3 comparison to the next lesson. Record the reduced coverage instead of compressing every activity into an unrealistic timetable.

Without student devices, print the source data and use the board for modelling. The teacher can prepare the plan with AI beforehand. Without a printer, display one clear record at a time and let students organise it in notebooks; check that every learner can access the display or supply an appropriate alternative.

If you want practical science work, start with an existing school-approved investigation and its actual procedure. Confirm equipment, supervision, material information, disposal, and learner arrangements through local processes. Then ask AI to improve the questions and evidence collection within that procedure. A generated experiment is a proposal requiring review, and a paper-data lesson remains a useful option when practical work cannot be supported.

For remote or catch-up work, provide the source material, one modelled example, numbered instructions, and an individual task. State how students submit their reasoning and where they can seek help. Do not give an absent student only the group's completed answer sheet; it hides the work you need them to do.

9. Review the AI draft with the source and answer key open

Read every assessment question and solve it yourself. Check the mathematics, passage references, property comparisons, and scoring decisions. AI can help locate possible mismatches, but its agreement with its own draft is not independent verification.

Audit the lesson against the supplied curriculum and source materials.
Do not rewrite it yet. For each substantive issue, return:
- The exact lesson passage or task.
- The requirement or evidence it conflicts with.
- The effect on teaching or assessment.
- The smallest useful correction.

Check content accuracy, objective coverage, answer keys, access,
individual evidence, timing including transitions, and feedback use.
Distinguish a verified mismatch from a question requiring teacher judgment.
Then revise only the affected sections and show any changed answers.

Curriculum and source materials: [paste]
Lesson and answer key: [paste]

Use a checklist to confirm that the plan includes the necessary materials, questions, and collection time. Review the answers and class-specific arrangements yourself; having every heading filled in does not establish that the lesson will work.

10. Save the teaching pack and use the student work to revise it

Save the teacher plan, student materials, and answer key as clearly separated sections or files. Put the curriculum source and lesson objective on the teacher page. Keep model answers away from student task pages. An editable document is useful for adapting the plan; a PDF can serve as a distribution copy after you inspect it.

Check the version students will actually use: table labels, page breaks, text size, symbols, and reading order. Make sure screenshots are not the only way to access essential data. Open links and confirm that a downloadable worksheet matches the lesson and answer key.

Record the class-specific decisions still required: selected text or data, timing adjustments, response routes, and any practical arrangements. A documentation check should not be labelled classroom testing. Keep a simple version date so colleagues can identify which plan and materials belong together.

After the lesson, sort individual work by the missing criterion. In the science example, separate accurate comparisons with weak atomic reasoning from inaccurate table reading. Give each group of responses a specific next task, then check the revised work.

The Education Endowment Foundation's feedback guidance recommends planning how pupils will receive and use feedback. Reserve time for that response. “Use more evidence” becomes actionable when students identify the missing comparison, add it to their explanation, and try a similar question.

Lesson revision note
Outcome and lesson version:
Evidence collected:
Criterion students met:
Criterion still missing:
Example of the issue, anonymised within approved school systems:
Next teaching move:
Task that will check the improvement:
Time or access adjustment for the next lesson:

Keep that revision note with the source materials and answer key. When you reuse the plan, adjust the part the student evidence showed was difficult and check whether the new class needs the same support.