NDIS Provider Worker Supervision and Competency Records: Use case and AI Confidentiality

Supervision is where a team leader writes down how a support worker is going: what he has been signed off to do, what went wrong and what happens next. The Human Resource Management quality indicators expect records of checks and training, timely supervision and documented performance, so a service keeps one for each worker, and team leaders ask ChatGPT or Claude to tidy them or to roll them up into NDIS compliance reporting for the board. This page marks every item in a supervision record for a casual worker after a complaint: his clearance, his sign-offs, a participant’s diagnosis, a paste into ChatGPT and his own circumstances. It shows where each item goes and how the same record is handled on Ermos.

The worker supervision and competency record

Written by the team leader at each supervision meeting and kept on the worker’s file. The Human Resource Management quality indicators expect records of pre-employment checks, qualifications and experience, a system to plan and record training, timely supervision, and worker performance to be managed, developed and documented. They also expect each worker’s contact details and any secondary employment to be recorded and kept up to date.

Brushbox Community Supports · Support worker supervision and competency record
Monday 9 November 2026, 4.00pm, at the office with Beth Hargreaves, Team Leader · Next supervision Monday 7 December 2026

Worker
Signed off
Rating key
What happened
The participant
The ChatGPT paste
Rhys’s circumstances
Agreed
Training
Signed

An invented supervision record. The clearance number and phone number are placeholders. Hover or tap any highlighted item for the analysis; the full list is below.

Supervision record drafter

Your team leader can use an Agent+ to draft the supervision record from their meeting notes and the worker's existing file, in your template. It carries across training and sign-off history, notes what was agreed, and marks any participant detail in the notes so it can be removed from the worker's record.

What it does

  1. Reads the meeting notes and the worker's file
  2. Drafts the record in your supervision template
  3. Flags out-of-date training, sign-offs and contact details

What stays with your people. The team leader reads, corrects and keeps the record, and discusses it with the worker. The agent drafts only.

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 service

  1. 1
    The worker, his clearance and his orientation certificate

    A named worker, his number, the clearance reference and expiry, when the database was checked and when he did the orientation module. The Human Resource Management indicators expect records of pre-employment checks, qualifications and experience and an induction that includes the mandatory NDIS worker orientation program. Section 7B(3) of the Privacy Act exempts an employer’s act directly related to a current or former employment relationship and an employee record it holds about the individual. That exemption is a separate matter from the indicators, which are about the records themselves. Privacy Act guide.

  2. 2
    What he is signed off to do

    Which supports Rhys may deliver alone and which he may not, including a second participant’s behaviour support plan and the existence of a PRN tablet. The Human Resource Management indicators expect the skills required of each position to be documented and training to be planned and recorded. The Medication Management standard applies to a provider responsible for administering medication; whether prompting a participant to take his own tablets counts is a question to settle against the wording of the standard. Privacy Act guide.

  3. 3
    The rating key

    The three-letter scale that appears on every worker record the service keeps. It names no one.

  4. 4
    The complaint and what was said

    A complaint reference, an allegation about a named worker and his own account. The Code of Conduct applies to persons employed or engaged by an NDIS provider and requires supports to be provided in a safe and competent manner, with care and skill, and requires them to promptly raise and act on concerns. Rhys raised this himself. An assessment that the matter is a Code concern is not a finding that he breached it. Privacy Act guide.

  5. 5
    A participant’s diagnosis and how to speak to him

    Declan’s diagnosis, his state of health that month and what works with him, written in another person’s file. Health information, which is sensitive information under the Privacy Act, and the Information Management standard expects each participant’s information to be confidential. The Privacy and Dignity standard expects supports that respect his dignity and privacy. Patient data guide.

  6. 6
    The paste into ChatGPT

    A worker put a chat holding a participant’s name, address, lockbox code, treatment and family into ChatGPT. APP 6.1 limits disclosure for another purpose, and APP 8.1 applies before an organisation discloses personal information to an overseas recipient. If an APP entity is aware of reasonable grounds to suspect there may have been an eligible data breach, it must carry out a reasonable and expeditious assessment and take all reasonable steps to finish it within 30 days of becoming aware (section 26WH); an eligible data breach involves unauthorised disclosure that a reasonable person would conclude is likely to result in serious harm (section 26WE). Whether this is one is for the service to assess. The Commission’s February 2026 position statement on AI says disclosing sensitive participant information to a third-party AI platform without safeguards may breach the Code, although it is about behaviour support plans. Privacy Act guide.

  7. 7
    The worker’s own health, family and second job

    His sleep, his father’s stroke, his night work and a rostering request. Health information about Rhys and about a relative. The indicators expect each worker’s contact details and secondary employment to be recorded and kept up to date, and the Privacy Act lists health information among the examples of personal information in an employee record. Whether sending it to an AI company counts as an act directly related to the employment relationship is a question for advice. Privacy Act guide.

  8. 8
    What was agreed

    Who Rhys works beside, which shifts he does not work alone and a review date. The Human Resource Management indicators expect performance to be managed, developed and documented, including through feedback and development opportunities. It also names Kirra Rennick and shows when a man who lives alone is, and is not, with a worker. Privacy Act guide.

  9. 9
    Training plan and certificate dates

    A session, a trainer, a first aid expiry date and a refresher. The indicators expect a system to identify, plan, facilitate and record training, with mandatory training identified, and infection prevention and control training, including refreshers, for all workers providing supports. Low sensitivity on its own, but it names a contractor and the dates of a worker’s certificates.

  10. 10
    Sign-off and review

    Who read it, who agreed it and who reviewed it. The names of two managers and the worker, with the date the worker accepted the record.

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 service 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 participant 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 Use cases in NDIS services.

Common questions

Do ChatGPT and Claude use the information in a worker supervision and competency record?
Yes. A record pasted into ChatGPT or Claude 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. One record carries a participant's schizoaffective disorder and what works when speaking to him, complaint C26-036 about a named casual worker, and that worker's sleep, night job and his father's stroke.
Can we use ChatGPT or Claude for a supervision record if we remove the names?
Not safely, because the record still describes two people without their names. 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 casual worker since March with a complaint on 31 October, a father's stroke and a warehouse night job is easy to place, and so is the man with a named diagnosis whom he was rostered to support. 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.
What if staff use their own ChatGPT or Claude account for a supervision record?
Then the service cannot see or manage what was sent from that account. 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. This record shows it: Rhys pasted a team chat naming Declan, his address and the lockbox code from his own phone on a free account, and after he deleted the conversation the service could not confirm what had happened to it.
How can an NDIS provider use AI for a worker supervision record 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 record is drafted from the worker's file and the service's own supervision template, and the team leader checks it and signs it. The service also needs a rule that staff must use it, because nothing stops a worker pasting into a personal phone.
What records do the NDIS Practice Standards expect for support workers?
The Human Resource Management quality indicators expect the skills and knowledge of each position to be documented, records of pre-employment checks, qualifications and experience to be kept, a system to plan and record training, timely supervision, and worker performance to be managed and documented. They also expect each worker's contact details and secondary employment to be recorded and kept up to date. The indicators must be taken into account when an applicant for registration or a registered provider is assessed against the standards.
Does the Privacy Act cover a support worker's employment record?
Section 7B(3) of the Privacy Act exempts an employer's act or practice that is directly related to a current or former employment relationship and an employee record it holds about the individual. The exemption concerns the worker's own record. It does not remove the Act from the service's handling of participant information, and whether it reaches participant information written into a worker's file is a question for advice.
Does the NDIS Code of Conduct apply to casual support workers?
Yes. Section 5(2) of the Code of Conduct Rules makes NDIS providers, persons employed or otherwise engaged by NDIS providers, and members of the key personnel of NDIS providers, Code-covered persons. The preamble says the Code applies whether or not the provider is registered. Section 6 requires them, among other things, to respect the privacy of people with disability and to provide supports in a safe and competent manner, with care and skill. Breaching the Code is a civil penalty matter under section 73V of the Act.

See Dominion answering from your own documents.

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

Book a demo