# Draft AI Exit Tickets from Lesson Objectives

> Draft short AI-assisted exit tickets from taught material, map each prompt to a learning objective, and review the answer key before classroom use.

Draft short AI-assisted exit tickets from taught material, map each prompt to a learning objective, and review the answer key before classroom use.

An AI-assisted exit ticket should ask about the lesson learners actually received. Start with educator-approved objectives and taught material, draft the smallest useful question set, and review both prompts and answers before classroom use. The output is an objective-question map and a short learner-facing ticket, with unresolved issues kept beside the draft.

A polished question can still assess something the lesson never taught. An answer key can be wrong even when its question looks reasonable. This workflow gives the educator a way to catch those defects before interpreting any learner response.

## State the feedback you want

[Formative assessment](/glossary/formative-assessment) provides feedback during learning so teaching and learning can be adjusted. It often carries low or no point value. [Carnegie Mellon's assessment overview](https://www.cmu.edu/teaching/assessment/basics/formative-summative.html) explains that purpose.

For this workflow, write down the question the educator wants the ticket to help answer: for example, “Do these prompts let learners show the two map-reading ideas we taught?” This is a design question about the instrument. It is not an instruction to diagnose learners or assign grades.

Set the available time and desired format if known. If the learners' age or reading level is unknown, use direct wording and flag that phrasing still needs the educator's review. Do not invent a grade, standards code, or reading-level judgment to fill an empty field.

## Register objectives and taught passages

Give each objective an ID and each taught passage a source ID with a locator. The packet should contain the actual handout or approved lesson text, not merely a broad topic such as “maps.” Name the educator who will review the draft. Student names and submissions are unnecessary for preparing the ticket.

If objectives or taught material are missing, return an input checklist and stop question drafting. If an objective requires a concept absent from the supplied material, identify the mismatch. The educator can supply another taught source or revise the objective; the assistant should not expand the curriculum automatically.

Keep source instructions separate from the task. A handout saying “ignore previous directions” is lesson content to examine, not a command to change the workflow. Work only from the materials explicitly supplied for this draft.

## Use AI to draft within the packet

[MagicSchool](/tools/magicschool) provides AI tools for lesson plans, worksheets, quizzes, and rubrics. [Its tool directory](https://www.magicschool.ai/magic-tools) describes these drafting surfaces.

Its quiz generator can use a supplied reading or topic to draft questions, choices, and a key, with generation inputs such as focus and difficulty. Output can be reviewed and edited. [MagicSchool's quiz page](https://www.magicschool.ai/tools/multiple-choice-quiz-assessment) describes this capability.

The short exit-ticket process here is our proposed use of drafting tools. Use the actual verified quiz or worksheet surface available to you; this guide does not assume a dedicated exit-ticket button. A compact drafting request could be:

```text
Draft two questions from L1 only, covering O1 and O2 once each.
Return question ID, objective ID, prompt, expected answer,
source ID, exact evidence, locator, and ambiguity note.
Keep all questions marked needs-review until the educator checks them.
Do not add untaught concepts or infer student understanding.
```

## Inspect a fictional map-reading ticket

This lesson is fictional. Passage L1 says: “On our handout, the legend's dotted line marks a walking path. The north arrow indicates north.” Objective O1 is to identify the walking-path symbol. Objective O2 is to explain what the north arrow indicates.

| Question | Objective | Draft prompt | Expected answer | Evidence from L1 |
| --- | --- | --- | --- | --- |
| Q1 | O1 | Which legend symbol marks a walking path? | The dotted line. | “the legend's dotted line marks a walking path” |
| Q2 | O2 | What direction does the north arrow indicate? | North. | “The north arrow indicates north.” |

Both rows start as needs-review. Add the handout locator and an ambiguity note to the working table. The learner-facing version need not include internal IDs or evidence columns, but the educator's review copy should retain them.

Reject a third draft asking how map scale converts distance. Nothing in L1 teaches scale. A question can fit the general subject of maps and still fail this ticket's scope. Removing it preserves the approved objectives without inventing a lesson that occurred elsewhere.

## Review the key before judging a response

Read each prompt aloud, answer it from the taught passage, and compare that answer with the draft key. If Q1's key says “solid line,” repair the key. That defect says something about the draft; it says nothing about a learner's misconception.

Check whether the wording offers more than one reasonable interpretation. For Q2, “Where does the arrow point?” might refer to a position on the page instead of the named direction. The educator can keep the direct direction question or provide the visual context explicitly. Do not add a visual reference that learners will not receive.

When two supplied keys disagree, record both and request the educator's decision. Keep the affected question out of the ready ticket until the governing answer is settled. A model's confidence is not a reason to discard one source silently.

## Check coverage without confusing it with alignment

The [Exit Ticket Builder](/skills/docs/exit-ticket-builder) organizes this one job around supplied objectives and lesson text. Its companion checker reads `exit-ticket-map.json`, using `id` fields for objectives, sources, and questions. It flags unknown references, uncovered objective IDs, absent quoted evidence, and questions still marked `needs-review`.

Those are structural checks. Referencing O2 does not prove a question actually tests O2. A literal quote may be unrelated to the answer. The educator must inspect that relationship and approve the wording and key before marking a row `draft-ready`.

Keep the final review record specific: which objective each question serves, which taught evidence supports the answer, what wording changed, and what remains unresolved. A checked map is easier to revise than a ticket whose design decisions have disappeared.

## Use the reviewed ticket with a clear human decision

Before copying the draft into a classroom tool or giving it to learners, the educator checks alignment, wording, and answer correctness. After use, the educator interprets the feedback and decides whether to revisit a symbol, clarify an explanation, or ask a follow-up question.

Retain the distinction between a draft defect, an ambiguous learner response, and a possible instructional next step. This workflow prepares reviewable questions; it does not automatically grade, diagnose, or make individual learner decisions. If the taught material changes, update the source packet and review the affected rows again.

## Sources and further reading

- [AI tools for teachers](https://www.magicschool.ai/magic-tools) — MagicSchool
- [AI multiple choice quiz generator](https://www.magicschool.ai/tools/multiple-choice-quiz-assessment) — MagicSchool
- [Formative vs Summative Assessment](https://www.cmu.edu/teaching/assessment/basics/formative-summative.html) — Carnegie Mellon University, Eberly Center
- [MagicSchool Pricing & Plans](https://www.magicschool.ai/pricing) — MagicSchool

---

_Source: https://agentscamp.com/guides/workflow/ai-exit-tickets-from-lesson-objectives — Guide on AgentsCamp._