Healthcare Patient Google Review Reply Emails: Use case and AI Confidentiality

A one star Google review lands on a Saturday and the practice manager wants a calm, professional reply by Monday. The quickest route is to paste the review and the patient’s chart notes into ChatGPT or Claude and ask for a response. The result is fluent, and it may confirm that the reviewer’s husband attended, what was found and what was prescribed, which the practice must not say in public. This page marks each item in the email that carries the review, the chart extract and the drafts, against the Privacy Act, Ahpra’s advertising rules on testimonials and Good medical practice, and shows how the same draft is handled on Ermos.

The email drafting a reply to a patient’s Google review

Sent by the practice manager to the doctors after a negative online review that mentions a clinical visit. It carries the review, the doctor’s summary of the consultation and the draft reply for approval.

From: Helen Tran
To: Dr Helena Voss; Dr Chidi Eze
Sent: Monday 2 November 2026, 4.10pm
Subject:

Helena, Chidi,

A review went up on Saturday morning. Here is what it says:

Chidi gave me this summary so I could answer properly:

I wanted something calm but kind, so It gave me this:

I have not posted it. It confirms he was our patient and says what we did and what the hospital found, which I do not think we can do. Here is the version I would post:

Two other points.

On the waiting time:

An invented email. The reviewer, her husband and the visit are fictional. Hover or tap any highlighted item for the analysis; the full list is below.

Online review reply drafter

An Agent+ that drafts a reply to a negative online review using the practice's own reply guidelines and tone. It works from the review text and the practice manager's notes, keeps the draft general, and flags any sentence that refers to a visit or to the patient's care.

What it does

  1. Reads the review and the practice's reply guidelines
  2. Drafts a short, general reply in the practice's tone
  3. Flags wording that refers to the patient's care

What stays with your people. The practice manager and the treating doctor approve the final wording and post the reply. The agent drafts; it does not publish.

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

  1. 1
    The reviewer’s name and the rating

    A Google display name and a one star rating. The review is already public, so the name is not secret. Tying it to a patient record is what turns a public comment into a privacy question for the practice.

  2. 2
    The review itself

    A public account of a clinical visit, written by the patient’s wife. The patient is free to say what he likes about his care. The RACGP says a practice must not breach a patient’s privacy no matter what the patient has posted, and should not confirm whether the reviewer attended or share why or what treatment was given. Patient data guide.

  3. 3
    The chart extract

    A name, date of birth, symptoms, an ECG result, a prescription, advice given and what happened at the hospital. All of it is health information (Privacy Act s 6FA) and so sensitive information under s 6(1), collected to provide a health service and used here for a different purpose, which engages APP 6.1. Patient data guide.

  4. 4
    Pasting the review and the chart into ChatGPT

    The step itself. The review and the summary go together to an AI company, which can link the public name to the patient’s health information. It is a disclosure by the practice, and an act of an employee in the course of their duties is treated as the practice’s act (s 8(1)(a)). Patient data guide.

  5. 5
    The AI draft that confirms attendance and treatment

    A reply that says he attended, what was done and what the hospital found. Posting it would disclose health information in public. It also confirms, in writing, that the reviewer’s husband was a patient, which the RACGP advises against even when the reviewer has posted details herself. Patient data guide.

  6. 6
    The reply to post

    Generic wording that thanks the reviewer, says the practice cannot discuss any person’s care in public and offers a direct conversation. It discloses nothing and matches what the RACGP suggests: keep public responses simple and take the conversation offline.

  7. 7
    The advertising rule on testimonials

    Ahpra’s testimonial tool defines a testimonial as recommendations or positive statements about the clinical aspects of a regulated health service used in advertising, and says advertisers should take care when engaging with reviews on third-party websites because that may be considered using a testimonial. Section 133 of the National Law is the provision it refers to. Advertisers are not responsible for removing testimonials on platforms they do not control.

  8. 8
    Asking Google to remove the review

    The RACGP says a practice can ask a site to remove a review that breaches the site’s policies or that it considers defamatory, that most sites will not remove a review because the practice disagrees, and that the practice should seek advice from colleagues, its medical defence organisation or a legal adviser before responding. The insurer’s terms decide what its policy covers.

  9. 9
    The waiting time and the receptionist

    Operational facts about a busy afternoon and a named receptionist. Ahpra says comments about customer service or communication style are not testimonials, so a general reply about waiting times can be made without breaching privacy. The staff detail is internal and is not for the public reply.

  10. 10
    The approval step

    The instruction that nothing is posted until a doctor approves. The decision about what the practice says in public stays with a person.

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 patient record. 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 Use cases in healthcare practices. See how this works in practice: Patient messages on Ermos.

Common questions

Do ChatGPT and Claude use the information in an email drafting a reply to a patient's Google review?
Everything pasted or uploaded 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. This email carries the review, Mr Cranston's date of birth, his chest tightness, normal ECG and prescription, and a draft reply that confirms he attended.
Can we use ChatGPT or Claude to reply to a patient's Google review if we remove the names?
Not safely, because the chart summary still describes one visit. 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 Thursday afternoon visit with two days of chest tightness, a normal ECG and an emergency department trip that night would be recognised by the reviewer, the doctor and the receptionist. 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. The RACGP advises a practice not to confirm that a reviewer attended, even when the reviewer has posted details herself.
What if staff use their own ChatGPT or Claude account to reply to a review?
The business cannot see or manage what was sent. 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. Both companies can also review content that is flagged for safety or policy reasons. Whatever the account, an act of an employee in the performance of their duties is treated as the practice's act under section 8(1)(a) of the Privacy Act, so a pasted chart summary is the practice's disclosure.
How can a practice use AI for a reply to a Google review 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 reply is drafted from the review and the practice's own reply template. Dr Voss reads every word before anything is posted, because a draft can still say too much.
Can a practice reply to a Google review?
The RACGP's guidance on responding to online reviews (page updated 29 June 2026) says the usual approach is to do nothing, and to respond only where a review raises a legitimate administrative or non-clinical point. Whatever the patient posted, a practice must not breach their privacy, and should not confirm whether the reviewer attended or share why they attended or what treatment they had.
Are patient reviews testimonials under Ahpra's advertising rules?
Ahpra's testimonial tool defines a testimonial as recommendations or positive statements about the clinical aspects of a regulated health service, used in advertising. Comments about customer service or communication style are not testimonials. Advertisers are not responsible for removing testimonials on platforms they do not control, but should take care when engaging with reviews on third-party sites.
Does the Privacy Act stop a practice confirming that someone is a patient?
Whether someone attended a practice is personal information the practice holds as a health service provider. APP 6.1 says information collected for one purpose must not be used or disclosed for another unless the person consents or an APP 6.2 exception applies. The RACGP advises practices not to confirm that a reviewer attended, even when the reviewer has posted details publicly.

See Dominion answering from your own documents.

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

Book a demo