Introduction
Loops AI enhances your data analysis workflow by helping you uncover insights faster. This guide will walk you through using our AI tools effectively to analyze survey data and extract meaningful patterns.
AI's Role in Your Analysis
Loops AI is designed to complement analysts, not replace them. Here's how the workflow typically unfolds:
AIs Role
Summarising large amounts of data at speed.
Doing complex analysis or breakdowns.
Prompting potential areas to explore further.
Analysts Role
Creating the assumptions or hypothesis.
Finding areas to explore further.
Creating the narrative of the analysis.
Checking the AIs results and presenting it in a digestible format.
Getting Started with Loops AI
Follow these steps to begin your AI-powered analysis:
Click the Loops AI button to open the interface
Select your AI model & data to include in your analysis
Choose specific questions you want to analyze
Add your data and craft your prompt to guide the AI
Choosing the Right AI Model
We offer three specialized AI models to suit different analysis needs:
Model | Best For |
Claude | Text analysis (comments) |
Gemini | Demographic analysis (audience splits) |
ChatGPT | Quantitative analysis (scores) |
To change models, click the dropdown menu in the AI configuration panel.
What Data to Include
Select what information to send to the AI for optimal results:
Demographic Data
Include when you need to explore differences between demographics
Exclude for general summaries to avoid hitting AI capacity limits
Answer Selection
Include to help the AI understand how respondents answered, connecting comments to scores
Comments
Include to analyze the content of what people have selected
Exclude if you're only interested in analyzing scores/selections
History
Include for iterative conversations with AI
Exclude for one-off, specific questions
Recommendation: Always include answers, comments, and history for most comprehensive analysis. Only add demographics when drilling down into specific segments.
Crafting Effective Prompts
Follow these steps to create prompts that generate the most useful insights:
Select your questions carefully
Be specific; don't select all questions unless necessary
Set the context
Start with looking at top themes rather than asking for general summaries
Specify which respondents you're interested in (e.g., all comments, only negative themes)
Clarify which questions relate to which concepts
Structure the output
Request specific formatting (e.g., themes with emojis, insights in tables)
Always ask for quotes to support themes or analysis
Iteratively refine your analysis
Choose specific themes to explore further
Provide context for why you want deeper insights
Tips for Getting Better Results
Be specific and detailed
Instead of "Give a summary of these comments," try "List the top 5 themes in the comments on X"
Structure your request clearly
Use formatting like bullet points or headers to organize instructions
Request supporting evidence
Ask for quotes that reflect and prove the AI's analysis (e.g., "Bring back 3 quotes for each theme")
Specify format preferences upfront
Mention your desired output format early (e.g., "Please respond with a table comparing each concept")
Use iterative refinement
Start with a basic prompt and refine based on responses
Sample Analysis Flow
Here's an example of how you might progress through an analysis:
Initial exploration: "What are the top themes for barriers to drinking Brand X in the last 6 months? Please list the top three themes with 3 quotes translated to English for each."
Focused analysis: "Please expand on the Taste/variant issues. What key themes do respondents mention? Bring back 3 quotes translated to English for each."
Specific deep-dive: "Analyzing comments on Taste/variant issues again, what do people say about which flavors they like/dislike? Please use quotes translated to English to support your answer."
Need Help?
If you have questions about using Loops AI for your analysis, please reach out to our support team.