Legal

AI Transparency Notice

Budgii PTY LTD (ABN 50 696 945 169 | ACN 696 945 169)

Effective Date: 24 April 2026 | Version 1.0

A plain, honest account of where artificial intelligence sits inside budgii and the rules we hold it to.

1. Purpose

This notice tells you where artificial intelligence is used inside budgii, what it does, what it does not do, how we keep it safe, and what rights you have in relation to AI-generated content or decisions. It is published in addition to our Privacy Policy and is written to meet the transparency expectations of the UK Information Commissioner’s Office, the Office of the Australian Information Commissioner, the New Zealand Office of the Privacy Commissioner, the United States Federal Trade Commission, and (for reference) the European Union Artificial Intelligence Act.

2. Where AI is used

AI is used in two distinct places inside budgii. Both are disclosed below. AI is not used anywhere else in the product, website, or customer communications without human review and sign-off.

  • The Nest Report. A monthly per-child report delivered to adult account holders, summarising behavioural patterns from the preceding month. Described in Section 3.
  • Content drafting assistance for articles published in the budgii Resources library. Described in Section 4 and in our Editorial Standards at budgii.io/editorial.

3. The Nest Report

3.1 What it is

The Nest Report is a plain-language monthly report on each Child in a household. It is produced by a large language model (“LLM”) using data recorded by the budgii app during the preceding calendar month, combined with development goals the Parent has selected at onboarding.

3.2 What it analyses

  • Scheduled vs completed to-do counts and rates.
  • Most and least reliably completed to-dos.
  • Current and longest Chain streaks, and days with zero completions.
  • Coins earned, spent, and saved.
  • Homework completion timing (on time, late, overdue).
  • App engagement (login days this month vs last month).
  • Age bracket and the Parent's selected development goals.

3.3 What it produces

  • A headline summary sentence and a sentiment badge (positive, mixed, or needs attention).
  • Up to five behavioural highlights tagged as wins, trends, or concerns.
  • Interpretations of chore and habit patterns, with an actionable recommendation for each.
  • Progress against the Parent's chosen development goals.
  • A focus area for the coming month and two to three practical micro-activities under ten minutes each.
  • A suggested conversation starter the Parent can use with the Child.

3.4 What it does not do

  • It does not diagnose any medical, psychological, or behavioural condition.
  • It does not use clinical language, and reports that do are blocked by the safety system described in Section 7.
  • It is not shown to the Child. Children never see AI-generated content about themselves inside budgii.
  • It is not used to make decisions that produce legal or similarly significant effects for anyone.
  • It is not shared outside the adult account holders for the household.

4. Content drafting assistance

AI is used to help draft articles published in the budgii Resources library. In every case, the angle, argument, and lived-experience grounding of the article are set by a member of the budgii team before drafting begins. A human reads and edits every article before it is published. No article is published unread.

The full description of this process is published in our Editorial Standards.

5. AI provider and data handling

TermMeaning
Primary AI providerAnthropic PBC, which provides the Claude family of large language models. Processing is carried out via Anthropic's API under a data processing addendum that contractually prohibits the use of API data for model training.
Processing locationAnthropic processes data in the United States. Data transferred from the UK, EU, Australia, and New Zealand is subject to appropriate safeguards, including the Standard Contractual Clauses and, where applicable, the UK Addendum.
Data sent to the modelOnly the behavioural data needed to generate the requested output. For Nest Reports, this is the Child's activity summary for the relevant month plus the Parent's selected development goals. No last name, location, school, image, voice, or device identifier is sent.
Data returned by the modelText content only. The output is passed through budgii's safety validation layer (described in Section 7) before being displayed to the Parent.
Retention of prompt and responseThe prompt sent to the model and the model response are retained for a short operational window to allow for error investigation, then deleted in line with our Data Retention & Deletion Policy.

6. Training and model improvement

No budgii data is used to train any third party’s AI model. Our agreement with Anthropic expressly prohibits the use of API data to train, fine-tune, or otherwise improve Anthropic’s models.

We do not sell, license, or otherwise make budgii data available to any third party for AI training purposes, and we do not operate any AI model of our own that is trained on user data.

7. Safety architecture

Every generated Nest Report passes through a safety validation layer before it is shown to the Parent. The layer performs the following checks:

  • Clinical-term block list. Reports that contain language commonly associated with clinical diagnoses (for example: ADHD, autism, anxiety disorder, depression, oppositional defiant disorder) are automatically blocked. A safe fallback report is shown in their place.
  • Harmful-content screen. Reports referencing religion, politics, medical advice, self-harm, violence, or other prohibited categories are blocked. A safe fallback report is shown.
  • Minimum data threshold. If a Child has fewer than seven active days in the month, a 'not enough data' report is shown instead of a full analysis.
  • Failure fallback. If AI generation or safety validation fails for any other reason, a generic fallback report is shown. The Parent is not shown an error message.
  • Rate limiting. Parents may request an out-of-cycle report manually, subject to a twenty-four hour cooldown per Child.
  • Human spot-check. A sample of generated reports is reviewed periodically by the budgii team for voice, accuracy, and safety compliance. Spot-checked reports are pseudonymised before review.
If a generated report would have breached any of the safety rules above, the Parent sees the fallback version, not the original. The original is not stored, not shown, and not delivered.

9. Your rights

You have the following rights in relation to AI-generated content:

  • Request a copy of any Nest Report generated about a Child in your household. Reports are already available in the app; this right extends to historical reports not currently displayed.
  • Request correction of any factual error in a Nest Report. We will review and, where appropriate, regenerate the report.
  • Request that a Nest Report be deleted. We will delete it from our systems and confirm deletion.
  • Object to the use of AI in generating Nest Reports. If you object, Nest Report generation can be disabled for your household at the account level.
  • Where applicable under UK GDPR Article 22 or equivalent laws, request human review of any decision. See Section 11.

Rights are exercised by emailing legal@budgii.io. We respond within thirty days in Australia, one month in the UK and EU, twenty working days in New Zealand, and forty-five days in the US (California).

10. Children's data and AI

We treat Children’s data sent to the AI provider as the most sensitive category of data we handle. Specific controls include:

  • No identifiers. The Child's name, age, family name, or any other directly identifying detail is either omitted or replaced with a household-internal placeholder before data leaves our systems.
  • Minimum data. Only the behavioural data required to generate the report is sent. Data from outside the report window is not included.
  • No training. As set out in Section 6, Children's data is never used to train any model.
  • Output safety layer. The safety architecture in Section 7 applies to Children's Nest Reports first and most strictly.
  • No Child-facing AI. Children never receive AI-generated content about themselves inside budgii.

11. Automated decision-making

The generation of a Nest Report involves automated processing, but we do not consider it a decision producing legal or similarly significant effects within the meaning of UK GDPR Article 22 or equivalent provisions. The report is an observational summary delivered to a Parent, who retains full human judgement about how to act on it.

Budgii does not use AI to make decisions about account creation, billing, suspension, content moderation, or any other decision that affects a user’s legal rights or access to the service. All such decisions are made or reviewed by a human.

12. Changes to this notice

We will update this notice when the way AI is used inside budgii changes. Material changes are notified to account holders by email at least fourteen days before they take effect. The effective date at the top of this page reflects the current version.

13. Contact

For questions about AI use, to exercise any of the rights in Section 9, or to request disabling of AI-generated content for your household:

legal@budgii.io

Budgii PTY LTD
ABN 50 696 945 169 | ACN 696 945 169
Sydney, New South Wales, Australia