Learning Timeline
Key Insights
Benefits of Code Execution
By enabling 'Code execution', the AI doesn't just guess data; it actually writes a Python script in the background to process raw data or perform precise calculations before presenting the table to you.
Source Transparency
The main advantage of this workflow is the ability to see every 'keyword search' performed by the AI during the 'multi-hop reasoning' process. This allows you to verify whether the AI is retrieving relevant information.
Prompts
Market Research on Coding Tools
Target:
Google AI Studio (Gemini)
Conduct a comprehensive analysis of the different AI coding tools currently available on the market.
Your task includes:
1. Perform multiple searches to find benchmarks, user reviews, and feature sets for top tools.
2. Use code execution to process any numerical data or rankings found.
3. Compile a detailed comparison table highlighting key differences.
4. Provide citations for every source used.
Step by Step
Configuring Gemini 3 Pro for Market Research
- Log in to Google AI Studio and open a new chat session.
- Select the 'Gemini 3 Pro' model (or the latest version supporting long context) from the model dropdown menu.
- Locate the 'Tools' section or system settings in the side panel.
- Check the box for 'Code execution' to enable the AI to write and run analysis code.
- Check the box for 'Grounding' (Google Search) so the AI can access live data from the internet.
- Enter a complex research prompt (see the 'Prompt Card' for examples) in the input field.
- Click the 'Run' button or submit the prompt.
Analyzing Multi-Step Reasoning Results
- Wait for the AI to finish generating the response; you will see the 'searching' and 'coding' processes happening in real-time.
- Scroll to the bottom of the output to view the generated Comparison Table.
- Click the dropdown or expand icon on the reasoning steps to see the list of search keywords used by the AI.
- Check the citations provided below the facts or table to verify source accuracy.
- If necessary, ask the AI to change the output format (e.g., to a slide format) based on the collected data.
