Reachy · Survey Guard Session
rohan@rivaltech.com
1. Upload & Map
2. Conversational Action
3. Results & Chat
I need to check the quality of my latest survey data. Can you help?
Reachy
I've opened Survey Guard for you. Upload your survey questionnaire and response data to get started. I'll help you through the analysis.
🛡 Survey Guard MCP App
1
Upload
2
Map Fields
3
Processing
4
Results
📄 luxury-brand-tracker-2026.docx 245 KB
📈 responses-wave3-n1204.csv 8.2 MB
Panel Type
Consumer
B2B
Paid Panel
Quick Poll
Reachy
I've opened Survey Guard for you. Upload your survey questionnaire and response data to get started.
🛡 Survey Guard MCP App
Upload
2
Map Fields
3
Processing
4
Results
Survey QuestionCSV ColumnConfidence
What is your opinion on luxury brands? Q1_LUXURY_OPINION 98%
How often do you purchase luxury items? Q2_PURCHASE_FREQ 95%
Which luxury brands do you prefer? Q3_BRAND_PREF 92%
Is sustainability important in luxury? Q4_SUSTAINABILITY 78%
+ 12 more mapped fields
Looks good, let's start the analysis
Reachy
Starting the analysis now.
🛡 Survey Guard MCP App
Upload
Map Fields
3
Processing
4
Results
67%
Analyzing response quality with AI...
Batch 8 of 12 · 804 of 1,204 participants evaluated
🛡 Survey Guard MCP App
Upload
Map Fields
Processing
4
Results
Total Participants
1,204
Critical Review
47
3.9%
Quality Concern
89
7.4%
Passed
1,068
88.7%
📄 critical-review-required.csv 47 rows ↓ Download
📄 quality-concern.csv 89 rows ↓ Download
📈 all-participants-consolidated.xlsx 1,204 rows ↓ Download
What are the main reasons people got flagged as critical?
Reachy
Based on the analysis, the 47 critical participants were flagged primarily for:

1. Straightlining (23 participants) - Gave identical responses across multiple matrix questions, particularly in the brand perception and sustainability sections.

2. Speed completion (14 participants) - Finished the survey in under 4 minutes (median was 14 minutes). Their open-ended responses were minimal or copy-pasted.

3. Gibberish/incoherent responses (10 participants) - Open-text answers contained random characters, repeated phrases, or content completely unrelated to the questions.

Would you like me to download the critical review file so you can examine the specific flagged responses?
Can you re-run with B2B panel criteria instead? The thresholds should be stricter.
Reachy
Switching to B2B panel configuration and restarting the analysis. B2B criteria use higher thresholds for thoughtfulness and relevance, which should catch more borderline cases.
+
🎤