Best interests file notes and AI confidentiality: what a broker’s record gives away in ChatGPT or Claude.

The best interests file note is the document a broker writes so that, years later, someone can see why this loan was recommended for this client. It records the client’s objectives, the lenders compared, the conflict check and the reasons, and it is the document most often drafted with help from ChatGPT or Claude because it is repetitive and everyone puts it off. It also carries the broker’s candid read of the clients. This page marks every item in a routine best interests note that the National Consumer Credit Protection Act and the Privacy Act bear on, and shows how the same note is drafted on Ermos inside the brokerage.

The best interests file note

Written on the day of the recommendation, from the file. Sections 158LA and 158LE of the National Consumer Credit Protection Act require a mortgage broker or its credit representative to act in the consumer’s best interests, and sections 158LB and 158LF require the consumer’s interests to take priority where there is a conflict. ASIC’s RG 273 sets out the records it expects brokers to keep.

File note: credit assistance
· Purchase and finance · A. Menon · 21 October 2026, 5.10pm

Objectives.

Circumstances.

Lenders compared.

Why Southgate.

Conflicts.

Inquiries and verification.

Impressions.

Recommendation.

An invented file note. The lenders are invented. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the brokerage

  1. 1
    The clients

    Named. Everything below is identifiable. Privacy Act guide.

  2. 2
    Objectives and requirements

    What the clients want from the loan, including a personal ambition. The inquiry into the consumer’s requirements and objectives that section 117 requires, written down. Privacy Act guide.

  3. 3
    Circumstances, worries and a family secret

    An employer’s lost tender, the client’s anxiety, a possible reduction in hours and a gift whose amount one parent has not been told. Confidences given to the broker to get the loan right. Privacy Act guide.

  4. 4
    The lenders compared

    The shortlist, rates and fees, and the lender that declined. Advertised rates are public; the choice of lenders for this household and the decline are not.

  5. 5
    Why this lender

    The broker’s reasoning, including what was explained to the clients. The core of the record ASIC’s RG 273.165 expects: the options, the recommendation and the reasons.

  6. 6
    The conflict check and the commission difference

    What each lender pays the brokerage and what the recommendation costs it. Commercial information about the brokerage, and the evidence for the priority rule in sections 158LB and 158LF.

  7. 7
    The record of the inquiries made

    The dates the credit guide was given, documents seen and the assessment completed. Low risk alone, and it sits beside the clients’ names. Section 113 requires the credit guide and sections 115 to 117 the assessment and inquiries.

  8. 8
    The broker’s impressions

    A candid read of the clients’ spending, their temperament and their risk. Written for the file, never for the clients. Privacy Act guide.

  9. 9
    The recommendation and the condition

    The loan, its structure and the step that makes the numbers work. The decision the clients are paying the broker to reach.

  10. 10
    The broker

    The author of the note, bound by the duty in section 158LA or 158LE.

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 brokerage 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 mortgage brokerages. See how this works in practice: Best interests notes on Ermos.

Common questions

Do ChatGPT and Claude use the information in a best interests file note?
A pasted note goes to the AI company's servers and is 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. This note holds the clients' objectives, a gift whose amount one parent has not been told and the broker's candid view that Daniel understates his spending.
Can we use ChatGPT or Claude for a best interests file note if we remove the names?
Not safely, because the detail in the note still points to one couple. 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. An employer that lost a tender in August, a $60,000 gift from a father, a purchase at Glenmore Park and a decline from one lender on 20 October fit one household. 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.
Can ChatGPT or Claude remember the details in a best interests file note later?
They can. Both products also have memory features that can carry details from one conversation into later ones, and on a personal account each user decides whether they are switched on (ChatGPT, Claude). A broker who drafts these notes in the same account, day after day, could see one couple's objectives, family confidence and candid impressions follow the user into a later chat about a different client. The clients gave those details to the broker to get one loan right.
How can a mortgage brokerage use AI for a best interests file note 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 note is drafted from the client file and the brokerage's own note template. The broker checks that it records the reasoning actually applied and signs it.
What records does ASIC expect a broker to keep?
RG 273.165 expects records such as the responsible lending assessment, the credit guide given, information provided to the lender, outcomes of applications, relevant conversations, the options and the recommendation with the reasons, and any conflict identified with the action taken. RG 273 is guidance rather than law, and it says keeping notes as the work is done may help capture the broker's reasoning.
What is the conflict priority rule?
Under section 158LB a licensee that knows or reasonably ought to know of a conflict between the consumer's interests and its own, its associates' or its representatives' must give priority to the consumer's interests. Section 158LF applies the same rule to credit representatives. ASIC's RG 273 says a broker with a conflict it cannot resolve in the consumer's favour must not provide the assistance.
Can AI draft the file note?
A tool can organise the broker's own facts, and the law does not prescribe the tool. The duty to act in the consumer's best interests, and the reasoning that shows it, stay with the broker. A note that reads well but records reasoning the broker did not actually apply is a worse record than a short honest one.

See Dominion answering from your own documents.

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

Book a demo