$ man icebreaker

GTM · Personalization

Icebreaker

The opening line of a cold email that proves you researched the person. A specific observation, recent activity, shared interest, or company signal. Not "I see you're in SaaS." More like "saw your post on signal-based SDR workflows last week."


why it matters

when I was an SDR sending 200 emails a day, I didn't have time for icebreakers. every email started the same. "Hey [name], I noticed you're a VP at [company]..." and they all got ignored. because everyone could tell I didn't actually look. now I automate the research part so the icebreaker is real. generic openers get deleted. specific icebreakers get replies. the first sentence is the filter.

how I use it

every campaign I build has an {icebreaker} merge field in email 1. I don't write the icebreakers manually anymore. I have two paths depending on scale. path 1: a Clay research prompt that pulls LinkedIn profiles, recent posts, and headline data, then generates 1-2 sentence icebreakers. the prompt embeds the full email body so the AI writes it to flow naturally. path 2: an Exa SDK script that crawls each company's domain for case studies, news, and thought leadership — then outputs a CSV of contextual icebreakers ready to merge. I ran this across 73 companies in one session. the script hits site:{domain} for case studies, then runs a separate news search with date filters, cleans the HTML boilerplate, and outputs structured research per company. 73 companies, 150KB of contextual intel, zero manual research. that's the difference between spending a week on icebreaker research and spending an afternoon writing the script that does it for you.


real use cases
use caseAutomated Icebreaker Generation at Scale
Problem

Needed personalized icebreakers for 73 companies before a major trade show. Manual research was not feasible at that volume.

Solution

Built two parallel paths - a Clay research prompt for LinkedIn-based icebreakers and an Exa SDK script that crawls each company domain for case studies, news, and thought leadership.

Result

73 companies got unique, contextual icebreakers. Zero manual research. The Exa script produced 150KB of contextual intel in one afternoon.

ClayExa SDKPythonCSV pipeline

related terms
Research PromptMerge FieldsVariableExa
GTM knowledge guideall terms →
ShawnOS.ai|theGTMOS.ai|theContentOS.ai