Books Commander

LLM Usage Disclaimer

Last Updated: April 22, 2025

1. Introduction

BooksCommander.com utilizes a Large Language Model (LLM) — specifically via ChatGPT — to interpret your instructions and determine the appropriate QuickBooks Online (QBO) API endpoints to fulfill your requests. While the LLM has been carefully configured and trained to operate safely and transparently, it may make mistakes or misinterpret input.

2. How the LLM Works

The LLM attempts to match your natural-language instructions with a supported QBO API action. It then issues that action through our secure backend using your session credentials. The LLM has been explicitly instructed to clearly notify you of any actions it intends to take before those actions are performed.

If you are unsure what the LLM is about to do, you should:

  • Ask it for further explanation or clarification
  • Tell it to pause and not perform the action
  • Modify your request to ensure precision

3. User Responsibility

You are ultimately responsible for all actions performed using your Books Commander session ID. By using this service, you acknowledge and accept the following responsibilities:

  • Always review the LLM's explanation of its intended action before approving it
  • Never proceed with a task if you do not understand what the LLM intends to do
  • Prompt the LLM to provide further explanation whenever needed
  • After any action is performed, immediately review your QBO books and records to verify accuracy

BooksCommander.com does not monitor or validate the actions performed by the LLM. All commands are executed based on your instructions, and the results should be verified by you or a qualified accounting professional.

4. Limitations

The LLM may not always choose the optimal API for a given task, and its understanding of your intent is based entirely on your input. Errors or omissions in instructions may lead to unintended outcomes.

You are encouraged to use caution and ask clarifying questions to ensure the system behaves as expected.

5. Best Practices for Using ChatGPT

To help protect your sensitive financial data, we strongly recommend the following ChatGPT usage best practices:

  • After completing a session with Books Commander, consider deleting the conversation from your ChatGPT history to clear your session ID from the immediate chat.
  • Use temporary chats when interacting with the GPT. As of this writing, temporary chats are not stored in history, do not affect memory, and are not used to train OpenAI models. OpenAI may retain a copy for up to 30 days for safety review.
  • If you don't want your data used to improve ChatGPT, go to User > Settings > Data Controls and toggle off "Improve the model for everyone".
  • Regularly review and delete any ChatGPT memories that may contain sensitive information via User > Personalization > Manage Memories.
  • Be aware that OpenAI's settings and UI options may change over time. Always double-check your privacy controls and security settings before interacting with sensitive data.

6. Contact

For questions or feedback regarding the behavior of the LLM, please contact us at support@bookscommander.com.