1. Introduction
Atlas ("Atlas," "we," "us," or "our") uses AI-powered features to provide personalized financial coaching, insights, and reports. This AI Data & Consent Policy explains what data is used by our AI features, how that data is processed, what safeguards are in place, and how you can control your AI experience. By enabling AI features within the app, you consent to the data practices described in this document.
2. Not Financial Advice
Atlas is not a licensed financial advisor, broker, fiduciary, or tax professional. The AI features in Atlas provide educational information based on general personal-finance principles such as the debt snowball method, budgeting fundamentals, and expense categorization. They do NOT constitute personalized investment, tax, legal, or financial planning advice.
For decisions involving material financial impact — including investments, tax planning, debt consolidation through third parties, bankruptcy, or estate planning — consult a qualified, licensed professional. Atlas, its operators, and its AI service providers disclaim liability for any financial outcomes resulting from reliance on AI-generated content within the app.
3. What AI Features We Offer
AI Financial Coach
A natural language chat interface that allows you to log expenses, record debt payments, ask for budget insights, and receive personalized debt elimination advice — all through conversational text. The AI coach understands 28+ intent types including expense logging, debt payments, budget management, and spending trend analysis.
AI-Generated Financial Reports
Weekly and monthly reports that analyze your spending patterns, project debt payoff timelines, identify top spending categories, and suggest areas where you could cut expenses. These reports are generated on-demand or on a schedule you configure, and are stored securely in your account.
Smart Categorization
When you log an expense via the AI coach, the system automatically categorizes it based on the description you provide (e.g., "Starbucks" is categorized as Food & Dining). This helps maintain consistent budget tracking without requiring you to manually select a category every time.
4. What Data Is Sent to AI
When you use AI features, the following types of data may be sent to our AI processing pipeline:
- Financial summaries: Aggregate debt balances, minimum payments, interest rates, budget allocations, and spending totals by category — never raw transaction-level data in bulk
- Chat messages: The text messages you send to the AI coach, along with recent conversation history for context continuity
- Expense descriptions: The descriptions you provide when logging expenses (e.g., "Groceries at Trader Joe's") for categorization purposes
We NEVER send the following to AI providers:
- Your name or email address
- Bank account numbers or credit card numbers
- Social security numbers or government-issued IDs
- Your Supabase user ID or authentication tokens
- Any other personally identifiable information (PII)
5. How Your Data Is Processed
Your data follows a secure processing pipeline designed to minimize exposure:
- Backend gateway: Your financial data is sent from the app to our secure backend (Supabase Edge Functions). The mobile app never communicates directly with AI providers.
- PII stripping: Our backend removes all personally identifiable information before forwarding any data to the AI provider. Only anonymized financial summaries and your chat text are included.
- AI provider: We process your messages through third-party large language model providers (currently OpenAI, accessed via OpenRouter) operating under strict data processing agreements with zero data retention. The provider receives only the sanitized data necessary to generate a response.
- Real-time processing: All AI processing occurs in real-time on a per-request basis. We do not perform batch processing of your historical financial data.
- Encryption: All communication between your device, our backend, and the AI provider is encrypted via TLS 1.3.
6. AI Provider Data Handling
We have selected AI providers that adhere to strict data handling standards:
- No data retention: OpenAI and OpenRouter do not retain your data after processing your request. Once a response is generated, the input data is discarded.
- No model training: Your data is never used to train, fine-tune, or improve AI models. Our API usage is explicitly excluded from training data pipelines.
- Zero retention policy: This zero-retention commitment is enforced contractually through our data processing agreements with AI providers.
- Data processor role: The AI provider acts solely as a data processor under our direction — not as an independent data controller. They process data only as instructed by our backend and have no independent rights to use your data.
7. Data Safeguards
We implement multiple layers of protection to ensure your data is handled safely throughout the AI pipeline:
- PII Sanitization: Names, email addresses, account numbers, and other identifying information are programmatically stripped from all data before it reaches the AI provider
- Scope Limitation: The AI is restricted to debt elimination, budgeting, and expense analysis topics only. It is designed to refuse general knowledge questions, medical advice, legal advice, and any topics outside personal finance
- Rate Limiting: AI usage is capped at 200 requests per user per day across all subscription tiers. This limit is enforced atomically at the database level to prevent abuse
- Prompt Injection Protection: Our backend validates all user input against 27+ known prompt injection patterns before forwarding to the AI provider. Malicious inputs are blocked at the input layer
- Content Filtering: AI responses are validated and sanitized before being displayed to you. Responses that fall outside expected formats or contain inappropriate content are filtered out
- Row-Level Security: All AI conversation history and generated reports are stored with Row-Level Security (RLS) policies, ensuring only you can access your own data
8. Your Rights & Controls
You have full control over how AI features interact with your data:
- Revoke consent: You can revoke AI consent at any time from Settings. This immediately disables all AI-powered features on your account.
- Delete conversation history: You can request deletion of all AI conversation history. Once deleted, this data cannot be recovered.
- View chat history: You can view your complete AI chat history within the app at any time to see exactly what data has been exchanged.
- What happens when you revoke consent: Disabling AI consent will turn off AI coaching, AI-generated financial reports, and smart categorization. You will still be able to manually categorize expenses.
- Manual features continue: All manual features — debt tracking, expense logging, budgeting, payment recording, and progress tracking — continue to work fully without AI enabled.
9. Children's Privacy
AI features within Atlas are not available to users under the age of 18. We do not knowingly process financial data through AI services for minors. If we become aware that a user under 18 has enabled AI features, we will disable AI access for that account and delete any associated AI conversation history.
10. Changes to This Policy
We may update this AI Data & Consent Policy from time to time as we add new AI features or change AI providers. We will notify you of any material changes through an in-app notification and by updating the "Last updated" date at the top of this page. If a change materially expands the scope of data sent to AI providers, we will request your renewed consent before applying the change. Your continued use of AI features after non-material changes constitutes acceptance of the updated policy.
11. Contact Us
If you have questions about this AI Data & Consent Policy, how your data is used by AI features, or wish to exercise any of your rights described above, contact us at:
privacy@atlas-budget.com