The questions schools ask first.

Yes. Data protection is the first thing we designed, not an afterthought. All student and report data lives on Australian infrastructure (AWS Sydney, with disaster recovery in Melbourne). Limited data is processed overseas for AI inference: the notes and draft text needed for each request are sent to Google Vertex AI, which processes them outside Australia. Data is encrypted in transit and at rest, access is scoped to your school, and every view and change is written to an immutable audit trail. There are no student accounts: students never log in and have no profiles; they exist only as names on their teacher’s class roster.

The full detail is on our security and privacy pages, both written for the people on your IT and compliance committees.

That’s the whole point. Lantern builds a Writing Profile from a teacher’s own past reports, capturing their openers, their warmth, the way they frame a next step, and carries that voice through every draft. It is not a “formal / friendly” tone dropdown: it is a readable profile of how that specific teacher writes, and different classes can carry different voices.

If a teacher has no past reports to hand, a short guided wizard captures their style from scratch in a few minutes. The profile rebuilds over time as their writing evolves.

Always. Lantern is a drafting and checking assistant, never the author of record. The teacher remains the author and the approver: nothing is finalised, exported or shared until they have read it and signed off. They can edit any draft freely, accept or reject every suggestion, and the final words are theirs.

Professional judgement stays with the professional. Lantern speeds up the typing and the rule-checking; it does not replace the teacher’s knowledge of the child.

Lantern drafts strictly from what the teacher gives it: their typed or dictated notes about that student, plus the school’s reporting rules. It does not pull in facts from anywhere else.

As an extra safeguard, the compliance check flags any claim in a draft that isn’t grounded in the teacher’s notes and surfaces it for their judgement. It is never silently rewritten or quietly accepted. The teacher decides whether the wording is fair and true before it goes anywhere.

Most of the setup is ours. Your school sends us its reporting policy and we set up your school’s rules in Lantern from it, in the background. Once that’s done, a teacher can be up and running in minutes: they make a class, add their students and give Lantern five past reports.

The five past reports are how Lantern learns the teacher’s voice, and it won’t draft for a class until that voice is there. Teachers paste the reports in or take them from a file of last year’s reports. Students’ names are replaced with a placeholder: the reports are anonymous examples of how the teacher writes and are never linked to a student. A teacher without five past reports can write five in Lantern instead, which checks them but drafts nothing.

Teachers sign in with their school Google or Microsoft account, or with an email address and password. There are no student accounts: students are names on a class roster. Teachers add them one at a time or upload a class list, which most schools can pull from their student information system or build in a spreadsheet.

Lantern does not connect directly to student information systems. Once a teacher finalises a report, they can copy it to paste into your reporting platform, or download it as PDF, Word (.docx) or plain text.

Tell us your environment in a demo and we’ll map the rollout to your school.

Yes. Lantern works from the teacher’s own observations and notes and produces them in the teacher’s own voice, checked against the school’s own reporting policy. The teacher reviews, edits and approves every comment, so the report remains an authentic, accountable professional judgement, the same standard expected of a report typed by hand.

Schools that aren’t ready for AI drafting can use Lantern for compliance checks only: see the next question.

Yes. A school can run Lantern as a compliance checker only, with no AI drafting. Teachers write their comments as they do now, wherever they do now. They paste one finished comment into Lantern, or a whole class’s comments at once, and Lantern checks each one against your school’s reporting rules, flags what it finds and suggests fixes for the teacher to accept or reject.

The check is a careful second reader, not a guarantee. It can miss things, so the teacher still reads every comment before it goes out.

Compliance checking on its own is offered as a separate package and, like the full product, quoted for each school. Contact us and we’ll put together a quote.

No. We do not use student information to train third-party AI models. Your reports, student notes and prompts are not used for model training and remain subject to contractual controls and deletion rights with our AI providers. Drafting and checking call Google Vertex AI directly, with no broker in between. Vertex AI processes that limited data outside Australia.

Pricing is based on the number of students and your school’s needs, so we quote each school individually. Contact us and we’ll put together a quote for your school.

We publish the list. Lantern relies on a small number of vetted third-party service providers for AI drafting, checking and hosting. Student and report data is hosted in Australia. Limited data is processed overseas for inference: our AI provider, Google Vertex AI, processes the notes and draft text for each request outside Australia. Student information is not used to train third-party AI models, and remains subject to contractual controls and deletion rights.

The current list, and what each one does, is on our third-party service providers page.

Your reports are yours. Teachers can copy any finalised report or download it as PDF, Word (.docx) or plain text at any time, and if you leave we provide an opportunity to retrieve your data before deleting it. Deleted data is purged from primary stores after a 30-day recovery window, information may remain in secure backup systems for a limited period before being permanently deleted, and every deletion is recorded in the audit trail.