Report season,
back in your hands.

01 · Your time

The end of the report-writing weekend.

The hard part of reports was never knowing your students. You do that every day. The hard part is the hours: turning a term of observations into thirty polished, policy-compliant paragraphs, one careful sentence at a time, on a Sunday night. Lantern clears the blank page and the repetitive groundwork, so the evening is yours again. You still read every word, but you start from a draft that already sounds like you, not from nothing.

• No more starting every comment from a blank page
• No more hunting for a fresh way to say the same thing
• No more last-minute checks against the reporting policy
• No more worrying your executive will find a hundred errors

02 · Your voice

It sounds like you, because it learned from you.

Drop in a folder of reports you’ve written before and Lantern distils them into a Writing Profile you can actually read: your openers, your warmth, the words you reach for, how you frame a next step. Every draft is written through that profile, so it reads as your work, not a tone preset. Teach a different year level with a different register? Each class can carry its own voice. No past reports handy? A five-minute wizard captures your style from scratch.

03 · Your day

Built for the way reports can actually get written.

Not in one heroic weekend: in stolen minutes, between classes, a handful of students at a time. Catch a thought about a child while it’s fresh, speak it into Lantern, and move on with your day. Prefer to type? Do that instead. Dictation is the shortcut, not the only way in.

Dictate between classes

Speak your notes in the corridor while they are fresh: a sentence about each child. Lantern listens and tidies; you type nothing.

One student at a time

Read the draft, nudge a word, hit Save and Next. Flow through the whole class in one sitting. Everything autosaves as you go.

Last semester, one click

What you wrote about each student before is right there, so you can build on it instead of starting from a blank page.

Take it with you when you're done

Finalise each report, then copy it straight into your reporting system or download it as PDF, Word (.docx) or plain text.

04 · Your rules

Your rules, checked before you see the draft.

You shouldn’t have to keep the reporting policy memorised. Lantern writes each draft from your notes to your character limit, then checks it against your school’s rules: prohibited and preferred phrasing, required structure, the forward-looking final sentence. When a draft breaks a rule, Lantern fixes it and checks it again, changing only the sentences the rule points at so the rest stays in your voice. If it can’t fix something, it tells you, and you fix or dismiss it before you sign the report off. Anything that genuinely needs your professional judgement is flagged for you, never quietly decided on your behalf.

• No subjective praise without evidenceschool rule
• 350–500 charactersYear 6 limit
• Forward-looking final sentenceEnglish department

05 · Your call

The honest answers to what you’re really asking.

You’re right to be sceptical of AI near your students’ reports. So here is exactly how Lantern handles the three things every teacher asks us first.

Will it actually sound like me?

Yes. That is the whole point. Lantern builds a Writing Profile from your own past reports: your openers, the words you reach for, the warmth and the way you frame a next step. It drafts in that voice, not a generic one. The first time you read a draft, you should recognise yourself.

Am I still in control?

Always. Every comment is a draft waiting for you. You read it, edit a word or rewrite a sentence, then approve it. Nothing leaves the system as final until you say so, and your name is on it because the words are yours. Lantern is the assistant; you are the author.

Could it invent things about a student?

It will not slip anything past you. If a phrase in a draft is not grounded in your notes, Lantern flags it as “worth a check” and asks for your judgement rather than presenting it as fact. A flagged claim is never quietly rewritten or removed: it sits in front of you until you keep or cut it.

The principle underneath all three: the teacher is the author and the final approver. Lantern drafts, checks and flags. You decide. Every edit and approval is recorded in an immutable audit trail, so the record always shows that a person signed off, not a model.

The honest comparison

Why not just use ChatGPT yourself?

Plenty of teachers already paste a few notes into a chatbot and tidy up what comes back. It can produce a paragraph. What it can’t do is get you through a whole class safely, in your voice, in a way your school can stand behind.

It sounds like the chatbot, not you

When every teacher reaches for the same tool, every comment starts to read the same way. Lantern writes through your own Writing Profile, so the voice stays yours.

It can’t count characters

Language models are notoriously bad at length, so comments come back over or under your limit. Lantern counts the characters itself and redrafts once if a comment runs over. If it still doesn’t fit, it’s flagged for you, never cut off mid-sentence.

It doesn’t know your policy

A chatbot has never read your reporting guidelines. Lantern checks every draft against your school’s prohibited and preferred phrasing and required structure, fixes what breaks them and checks again. Anything it can’t fix is flagged for you.

It loses the class and the history

Pasting one student at a time loses last semester’s comment and the rest of your cohort. Lantern keeps the whole class set and each student’s history together. Each finalised report copies or downloads as PDF, Word (.docx) or plain text.

And the quiet risk: the moment you paste a student’s name and details into a free chatbot, identifiable student information has gone to a service your school has no agreement with, and many consumer tools reserve the right to train on it. Under the Australian Privacy Principles that obligation sits with your school. Lantern stores every draft and note in AWS Sydney, sends only the limited data each request needs to Google Vertex AI (which processes it outside Australia), and does not use student information to train third-party AI models.