Service agreements and AI data leakage: what a migration agent’s fee agreement gives away in ChatGPT or Claude.

The service agreement is the first document a registered migration agent signs with a client, and the Code of Conduct says what it must contain: the client’s name, date of birth, email and residential address, the agent’s MARN, the work to be done, how the file is handled and the fees. Staff paste it into ChatGPT or Claude to check the wording or turn it into a fee estimate email. This page marks every item in a routine service agreement that the Code of Conduct and the Privacy Act bear on, including the AI consent clause, and shows how the same agreement is handled on Ermos inside the practice.

The service agreement and fee estimate

Signed before any immigration assistance is given. Section 42 of the Code of Conduct requires a written service agreement signed by the client and the agent, covering the work, with the matters in section 42(3) and Division 3 (fees and disbursements) included. Section 43 allows advice at an initial consultation without one.

Service agreement and fee estimate
Kingfisher Lane Migration, Suite 3.07, 41 Aldwych Street, Parramatta NSW · 5 October 2026

Registered migration agent
Client
Work to be done
Fees
Disbursements
Consumer guide
Your file
Use of AI tools
Signed

An invented service agreement. The MARN is a placeholder and the practice, agent and client do not exist. The fees are the invented practice’s. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the practice

  1. 1
    The client’s identity block

    Name, date of birth, email and residential address, the details section 42(3) of the Code requires in every service agreement. Personal information, and the key that ties everything else on the file to one real person. Privacy Act guide.

  2. 2
    The agent’s name, MARN and contact details

    The Code requires them in the agreement. They are details a practice shows to attract clients, so this item adds little exposure on its own.

  3. 3
    The visa strategy and the expiry date

    What the client is trying to achieve and the date her current visa ends. Her whole risk in one line, and the work the practice is being paid to do. Privacy Act guide.

  4. 4
    Fixed fees, by stage

    Section 46 of the Code requires a fixed fee or an hourly rate with an estimate, including GST, and says it must be fair and reasonable. Commercial information about the practice and about what this client agreed to pay.

  5. 5
    The AI consent clause and the box ticked

    OMARA’s AI guidance says an agent needs the client’s written consent before entering personal details into an AI platform. The clause records the consent, but it says nothing about where the data goes once it is entered. Privacy Act guide.

  6. 6
    Standard terms: disbursements, consumer guide and file handling

    Template wording that sections 38, 42(3)(e), 47 and 54 of the Code require in substance, with no client data in it. The same text appears in every agreement the practice signs.

  7. 7
    Signatures and signing dates

    Who signed and when. A signed record the practice must keep on the client file for 7 years under section 56 of the Code.

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 EdgeErmos Dominion
Where the document is processedProvider servers, generally offshore on personal plansYour instance in an Australian data centre; answers generated in AustraliaThe unit in your office
Who can read itProvider staff and systems, under policyYour staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothingYour staff only
Is it keptOn the provider’s termsKept on your instance; the provider is barred from retaining itStays on the unit
Can it train a modelOn consumer tiers, yesNo, by contractOnly your own unit learns from it, on your hardware

Part of AI data leakage in migration practices.

Common questions

Do ChatGPT and Claude use the information in a service agreement and fee estimate?
Yes, whatever is pasted from a service agreement is sent to the AI company's servers and processed there to produce a reply. As at October 2026, whether a conversation on a personal plan (ChatGPT Free, Plus or Pro; Claude Free, Pro or Max) can be used to train the models depends on a setting each user controls. Business plans (ChatGPT Business and Enterprise, Claude Team and Enterprise) and the developer APIs sit under commercial terms, and OpenAI and Anthropic each say they do not train on that content by default (OpenAI, Anthropic). On either product, the thumbs up and thumbs down buttons can send a whole conversation back for training. The agreement holds the client's birth date and home address, her subclass 485 expiry of 19 February 2027 and the fixed fees for both stages of her application.
Is it safe to paste a service agreement into ChatGPT or Claude if we take the client's name out?
No, because the rest of the agreement still points to one person. Removing names rarely makes a document anonymous. Information stays personal information while the person is reasonably identifiable, and the OAIC warns that information can be at risk of re-identification even when it has been de-identified or anonymised. A birth date, a North Sydney address, a visa expiring on 19 February 2027 and a plan for a NSW nomination fit only one client. The Office of the Australian Information Commissioner (OAIC) recommends in its guidance on commercially available AI products that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools, because of the significant and complex privacy risks.
Does a ChatGPT Business or Claude Team plan make it safe for a service agreement?
Not on its own. A business plan changes the default on training, but it does not change who receives the content. On a personal plan the user cannot choose where conversations are stored or processed. Some business plans let an organisation choose to store data at rest in Australia (OpenAI), which still leaves the content with the AI company. The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product. The ticked consent clause records that the client agreed to AI drafting tools, yet it names no AI company, so the practice should check what she actually agreed to. Our comparisons with ChatGPT Enterprise and Claude weigh the trade-offs.
How can a migration practice use AI for a service agreement without sending it to ChatGPT or Claude?
Ermos is built so the document stays under your control. On the Dominion AI unit it is drafted inside your office and none of it is sent to an AI provider. On Ermos Edge your documents and index stay on your own instance in an Australian data centre, and only the question and the excerpts the user is permitted to see go to a named Australian provider that is contractually barred from keeping them or training any model on them. The agreement is drafted from the client file and the practice's own service agreement template. A registered migration agent checks every term and signs it.
What must a migration agent's service agreement contain?
Section 42 of the Code of Conduct requires it to be in writing and signed by the client and an agent in the business. It must cover the work, give each client's name, date of birth, email and residential address, name the agents and their MARNs, state that the consumer guide was given, explain what happens to the file, and set out fees and disbursements.
Does the AI consent clause make it safe to use ChatGPT or Claude on the file?
It addresses the written consent the OMARA says an agent needs, but consent does not change where the data goes. The client agreed to the practice using an AI tool; she did not agree to any particular provider's servers, retention or training terms. The agent also remains responsible for everything the tool produces.
How long must the practice keep the file after the agreement ends?
Section 56 of the Code requires a migration agent to take all reasonable steps to keep the client file for 7 years after the last action on it. The file must include the service agreement, its variations and the written communications. Anything pasted into an AI tool in that time is a copy the practice no longer controls.

See Dominion answering from your own documents.

A demonstration takes 30 minutes, uses no client data, and comes with no obligation.

Book a demo