AI Safety and Best Practices
Important guidance on using AI features safely and effectively in Tetri Copilot.
Overview
AI features in Tetri Copilot are powerful tools, but like any AI system, they have limitations. Understanding these helps you use AI effectively and avoid common mistakes.
AI Limitations
AI responses are based on patterns in your data and general knowledge. The AI can make mistakes, especially with: unusual or complex transactions, incomplete data, rapidly changing regulations, and highly specific legal or tax questions. Always review AI output before acting on it.
Best Practices
| Field | Required | Description |
|---|---|---|
Always Review | โ | Treat AI suggestions as a first draft, not a final answer. Human review is always necessary. |
Use Specific Questions | โ | Specific questions get better answers than vague ones. |
Verify Compliance Guidance | โ | Cross-check compliance advice with HMRC.gov.uk or a qualified accountant. |
Check AI Actions Before Approval | โ | Before approving an AI Action, confirm you understand exactly what it will do. |
Report Inaccuracies | โ | If the AI gives wrong information, correct it and note it for the team. |
Important
AI in Tetri Copilot is designed to assist, not replace, human judgment. For financial decisions, legal questions, and compliance filings, always involve a qualified professional.
Was this article helpful? ยท