Migration Agent Genuine Student Statements: Use case and AI Confidentiality
A student visa applicant has to show she is a genuine student, and the online form asks four questions about her family, work, money, course, plans and visa history, with a 150 word limit on each answer. Many applicants and the agents who help them ask ChatGPT or Claude to tighten the wording or cut an answer to length. The draft holds her mother’s health, her salary, an earlier refusal and what she told the agent in confidence. This page marks every item in a routine set of Genuine Student answers that the Code of Conduct and the Privacy Act bear on, and shows how the same draft is checked on Ermos inside the practice.
The genuine student statement
Prepared after the intake call and before a student visa application is lodged. Home Affairs says the Genuine Student requirement applies to applications lodged on or after 23 March 2024, prefers the answers in the online form to a separate statement, allows 150 words per response and gives more weight to statements supported by evidence.
Genuine Student answers, draft 2 for client review
Kingfisher Lane Migration · Monday 9 November 2026 · prepared by Tomas Eriksen for Renata Kowalczyk-Bell
| Applicant | Ruvini Karunaratne, born 2 April 2002, Sri Lankan citizen, passport N0000000 (placeholder), ruvini.karunaratne@example.com1 |
| Course | Master of Data Science, Kesterton Graduate College, Sydney, CRICOS 00000A (placeholder), starting Semester 1 20275 |
| Question 1: current circumstances (150 words, in English)8 | I live in Colombo with my mother, Kamala, a retired schoolteacher who has Type 2 diabetes, and my brother Kasun, who finishes school next year. My father died in 2021.2 I have worked as a software tester at Rangiri Systems (Pvt) Ltd since July 2021 and earn LKR 285,000 a month, part of which pays for my mother’s treatment. The family home in Nugegoda is in my mother’s name. My mother’s fixed deposits and my savings will pay for the course.3 My uncle Rohan has lived in Auburn, NSW, since 2009. I will rent my own room and will not depend on him for money. |
| Question 2: why this course, provider and location8 | I have tested payment software for five years and now spend most of my time on data quality. I want to become a data analyst. I compared Master of Data Science courses in Sri Lanka, Singapore and Australia. The Sri Lankan courses I found do not include an industry project. Kesterton Graduate College offers a two-year course with a six-month industry placement, and I have read the course handbook and the unit outlines for machine learning and data engineering. I chose Sydney because I can visit my uncle, and I have budgeted for rent, transport and health cover from the college’s cost guide. |
| Question 3: how the course will benefit you8 | A Master’s degree and the industry placement will let me move from testing into analysis. Rangiri Systems has told me in writing that it will hold a data analyst position for me when I finish, and the company is expanding its payments work in Colombo. I expect to return to Colombo after the course.6 Analysts with this training are in short supply there. My reason for coming is to study. |
| Question 4: any other relevant information8 | In 2022 I applied for an Australian visitor visa to see my uncle. It was refused because the bank statements I gave were incomplete. I have not applied for any other visa for Australia. I have never been refused a visa for any other country. My only overseas trip was to the Maldives in 2019 on a holiday.4 |
| Documents to upload | Rangiri payslips July 2025 to October 2026, an employment letter, fixed deposit certificates, BSc in Information Technology transcript (2023, studied part time while working), offer of place. |
| Agent’s private note (never goes on the form) | On the 26 October call Ruvini said she might apply for permanent residence “if it works out” and that her uncle has offered her a job in his cafe. The draft says she expects to return. Ask her what is true and change the answer to match.7 |
| AI use | Drafted with an AI drafting tool. Written consent signed by Ruvini on 26 October 2026 and held on the client file.9 |
An invented set of answers. The applicant, her family, her employer, the college and the CRICOS number do not exist, and the passport number and MARN are placeholders. The question headings summarise those Home Affairs publishes. Hover or tap any highlighted item for the analysis; the full list is below.
Genuine student answers drafter
An Agent+ that works from the intake call notes, the client's study history and the supporting documents. It drafts each answer in plain words in your template and marks the claims that need evidence. The student's own circumstances and intentions stay theirs to confirm.
What it does
- Reads the intake notes, study history and documents
- Drafts each answer within the word limit
- Marks claims that still need supporting evidence
What stays with your people. The student confirms every answer is true and their own. A registered migration agent reviews before anything goes into the application.
An Agent+ is built once and anyone in the practice can run it. It works from your own documents and runs on Ermos Dominion, the AI unit in your office, or on Ermos Edge, your own instance in an Australian data centre.
What leaves the practice
- 1The applicant’s identity block
Name, date of birth, nationality, passport number and email. Personal information, and the key that ties every answer below to one real person. Section 35 of the Code stops the agent disclosing it to a third person without the client’s written consent. Privacy Act guide.
Frontier Models (ChatGPT, Claude) Name, birth date, passport number and email transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 2Her mother’s health and the family
A named mother with a medical condition, a brother still at school and a father who has died. Health information about a person who is not the client is sensitive information where the Privacy Act applies, and the family detail is the client’s own account of her ties to home. Patient data guide.
Frontier Models (ChatGPT, Claude) A mother’s diabetes and the family transmitted; on consumer tiers they can train future models Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 3Her employer, salary and the family’s money
A named employer, how long she has worked there, her pay, who owns the family home and where the course money comes from. Employment and financial information. Home Affairs lists employment and economic circumstances among the matters it looks at and gives more weight to statements backed by evidence, so these figures must match the documents. Privacy Act guide.
Frontier Models (ChatGPT, Claude) Employer, salary and the family’s money transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 4The earlier visa refusal and travel history
A visitor visa refused in 2022 and the reason given. Clause 500.212 of Schedule 2 to the Migration Regulations 1994 has the Department consider the applicant’s immigration history. The answer must be accurate: section 234(1) of the Migration Act covers, in connection with a visa application, presenting a false document, making a statement known to be false or misleading in a material particular and delivering a document containing false or misleading information, and section 15 of the Code bars the agent from making a false or misleading statement.
Frontier Models (ChatGPT, Claude) Refusal and travel history transmitted, and the model can smooth the wording into something inaccurate Ermos Edge Stays on your instance, but the agent still checks every word against the documents Ermos Dominion Stays on the unit, but the agent still checks every word against the documents Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 5The course, the college and the uncle in Auburn
The named provider and course, and a relative in Sydney with his suburb. The Department looks at how much the applicant knows about the course and about living in Australia. The uncle is a third person who has not agreed to appear in an AI tool. Privacy Act guide.
Frontier Models (ChatGPT, Claude) Course, college and an uncle’s suburb transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 6A promise of a job and a statement of intent
The employer’s written promise of a role and the claim that she will return home. Section 20(2) of the Code bars the agent making a statement to a government official that the agent suspects, or ought reasonably suspect, is false or misleading without taking reasonable steps to check it. A model will happily make an intention sound firmer than the client holds it. If the employer’s letter is lodged, section 20(7) also applies to its origin.
Frontier Models (ChatGPT, Claude) The job promise and her stated intention transmitted, and the model can overstate both Ermos Edge Stays on your instance, but a model can still overstate an intention; the agent checks it with the client Ermos Dominion Stays on the unit, but a model can still overstate an intention; the agent checks it with the client Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 7The agent’s private note
What the client said about permanent residence and her uncle’s job offer, and the agent’s own instruction. It records an oral communication, which section 55 of the Code requires the agent to write down at the time, and it belongs on the client file for 7 years under section 56. Home Affairs says a later intention to apply for permanent residence is not counted against an applicant, so the note is sensitive for its candour rather than because the intention is a bar.
Frontier Models (ChatGPT, Claude) The agent’s candid note about her intentions transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 8The form questions and the word limit
The four Genuine Student questions and the 150 word limit are published on the Home Affairs website. Nothing about this client is in them.
Frontier Models (ChatGPT, Claude) Questions and word limit published by Home Affairs; harmless to send Ermos Edge Nothing to keep in; the questions are public Ermos Dominion Nothing to keep in; the questions are public Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case - 9The AI consent line
OMARA’s guidance says an agent needs the client’s written consent before entering her personal details into an AI platform, and should explain the use of AI in advance. The line records that consent. It does not say which tool, where its servers are or what it keeps. Privacy Act guide.
Frontier Models (ChatGPT, Claude) The line records consent, but not which tool is used or where its servers are Ermos Edge The consent is still needed whichever tool is used Ermos Dominion The consent is still needed whichever tool is used Use case Genuine student answers drafter. Drafts a first pass at each genuine student answer from the intake notes and documents, and checks each against the word limit and the evidence on file. See the use case
What happens in Frontier Models (ChatGPT, Claude)
- The whole document is transmitted to the provider’s servers. On personal plans the user cannot choose where it is stored or processed, and for the major providers that has generally meant outside Australia, under that country’s legal process. Some business plans offer storage in Australia, but the provider still receives the content.
- On consumer tiers, prompts can be retained and used to train future models.
- The provider’s staff and systems can reach the content under the provider’s own policies.
- Once sent, the practice no longer controls it. Retention and deletion happen on the provider’s terms.
The same document on Ermos
Ermos Dominion
The document is drafted on a unit in your office, from your own templates and the client file. Every marked item stays on the unit, which keeps working with the internet switched off. Nothing in it is transmitted to an AI provider, so there is no disclosure to analyse: the confidentiality work is the same as for your file server.
Ermos Edge
Your documents, the index and your access boundaries stay on your own instance in an Australian data centre. When a staff member asks for this draft, only their question and the excerpts they are permitted to see go to a named Australian provider, who is contractually barred from keeping the content or training any model on it. Nothing is processed outside Australia.
| Frontier Models (ChatGPT, Claude) | Ermos Edge | Ermos Dominion | |
|---|---|---|---|
| Where the document is processed | Provider servers, generally offshore on personal plans | Your instance in an Australian data centre; answers generated in Australia | The unit in your office |
| Who can read it | Provider staff and systems, under policy | Your staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothing | Your staff only |
| Is it kept | On the provider’s terms | Kept on your instance; the provider is barred from retaining it | Stays on the unit |
| Can it train a model | On consumer tiers, yes | No, by contract | Only your own unit learns from it, on your hardware |
Part of Use cases in migration practices.
Common questions
Do ChatGPT and Claude use the information in a Genuine Student statement?
Is it safe to paste Genuine Student answers into ChatGPT or Claude if we remove the applicant's name?
Can ChatGPT or Claude be trusted to draft Genuine Student answers?
How can a migration practice use AI for a Genuine Student statement without sending it to ChatGPT or Claude?
What does the Genuine Student requirement ask?
Can an AI tool write the answers for an applicant?
Does the agent need the student's consent before using ChatGPT or Claude?
See Dominion answering from your own documents.
A demonstration takes 30 minutes, uses no client data, and comes with no obligation.
Book a demo