If you’re working with a sensitive accountant or need to mitigate costs when using an AI / LLM, here are some strategies a user could employ:
- Be Concise:
- Keep your messages as brief and to the point as possible.
- Avoid unnecessary details or repetition.
- Use Shorter Responses:
- Request shorter answers when possible. For example, “Please provide a brief summary” or “Keep your response under 100 words.”
- Break Up Large Queries:
- Instead of asking one long, complex question, break it into smaller, focused questions.
- Avoid Repetition:
- Don’t repeat information from previous messages unless necessary.
- Reference previous points instead of restating them.
- Use Efficient Prompting:
- Learn to phrase your questions in ways that elicit direct, efficient responses.
- Be specific about what you need to avoid unnecessary elaboration.
- Limit Conversation History:
- Start new conversations for new topics rather than continuing very long threads.
- Batch Questions:
- If you have multiple related questions, ask them together to reduce the number of separate interactions.
- Use Formatting:
- Use bullet points or numbered lists to organize multiple questions or points, making it easier for the AI to provide structured, concise responses.
- Avoid Open-ended Queries:
- Instead of “Tell me about X,” ask “What are the three main points about X?”
- Review and Edit:
- Before sending, review your message to remove any unnecessary words or phrases.
By implementing these strategies, you can help reduce the number of tokens used in your interactions, potentially lowering costs while still getting the information you need.