Learning Timeline
Key Insights
The Importance of Block Structure vs. Standard Summaries
Don't just ask for a ‘summary’. By breaking the prompt into specific blocks like 'Uncomfortable Truths', you force the AI to think critically and provide deeper value than a standard text summary.
JSON Troubleshooting Tips
Ensure your JSON prompt doesn't have missing commas or unclosed brackets. A broken structure will cause the API call to fail (Error 400).
Prompts
Deep Analysis Prompt Structure
Target:
GPT-4 / GPT-3.5
Analyze the following transcription and provide the output in these specific blocks:
1. Uncomfortable Truths: Realities that are hard to face but necessary.
2. Action Items: Clear, immediate steps to take.
3. Long-term High Impact Opportunities: Strategic moves for future growth.
Transcription: {{transcription_data}}
Step by Step
How to Configure OpenAI API JSON Prompts in Make
- Open your Make scenario and select either the HTTP module or the OpenAI 'Create a Compilation/Chat Completion' module.
- Click on the 'Body' or 'JSON' field to start building your data structure.
- Insert the 'system' JSON object at the top to define the AI's role.
- Draft 'Instruction Blocks' within the prompt to enforce specific outputs (e.g., Uncomfortable Truths, Action Items, High Impact Opportunities).
- Use the 'Transcription' variable or input data from the previous module and place it under the system prompt instructions as context.
- Ensure the JSON structure is properly formatted so the API can parse each information category separately.
- Enter your API Key in the 'Connection' section of the OpenAI module.
- Map the previously built JSON variables into the fields required by the API.
- Save your configuration by clicking the 'OK' button.
