Data Downloads

All files are available for download below.

Files

File	Description
master_summary.pdf	Full analysis report (PDF)
master.csv	One row per conversation — 199 rows × 89 columns, UTC + Tehran timestamps
issues.csv	One row per LLM-extracted issue — 608 rows
all-conversation-pages.zip	All 199 individual conversation pages (Markdown)

Import to Google Sheets

1. Click the file link above and download
2. Open Google Sheets → File → Import → Upload
3. Select the CSV → separator: Comma → click Import data

Column reference — master.csv

Column	Description
conversation_id	Intercom conversation ID
app	App the conversation relates to (Notify Me!, Discounty, Subi, Convi)
shopify_domain	Merchant's Shopify store domain
created_at_utc / created_at_teh	Conversation start — UTC and Tehran time
closed_at_utc / closed_at_teh	Conversation close time
rating	Review star rating (all 5 in this dataset)
review_text	Full text of the review
sentiment	LLM sentiment (very_positive … very_negative)
resolution_status	LLM resolution assessment
install_to_conv_seconds	Seconds from app install to conversation start
install_to_conv_label	Human-readable install age bucket
seconds_to_human_handoff	Wait time before first human agent joined
seconds_from_handoff_to_review_ask	Time from human join to review request
fin_participated	Whether Fin AI bot was active
fin_resolution_state	Fin's outcome (Escalated / Resolved / etc.)
channel	How conversation started (customer_initiated / automated / admin_initiated)
region	Merchant country / region
contact_email_type	Email classification (personal_business / personal_free / team_mailbox)
agent_count	Number of unique human agents involved
blockers	Product blockers / missing features mentioned (semicolon-separated)
competitors_mentioned	Competitors named in the conversation