Everyone wants AI to write their cold emails. Almost nobody fixes the step before it. I built an ...Everyone wants AI to write their cold emails. Almost nobody fixes the step before it. I built an ...
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Everyone wants AI to write their cold emails. Almost nobody fixes the step before it.
I built an outbound engine in Make that enriches scraped leads, writes personalised follow-ups with GPT and syncs everything into Lemlist and Instantly. The GPT part took an afternoon. The other 80 percent of the build was unglamorous plumbing.
De-duplicating the scraped list before it touches anything else.
A router that parks the non-fits instead of emailing them.
Normalising enrichment into structured JSON so the model is not guessing.
Error handlers on every API call, so one bad record does not kill the whole run.
Status written back to the sheet, so the run is auditable afterwards.
Personalisation quality is a data problem wearing a copywriting costume. Feed a model thin context and you get I loved your recent post, at scale.
If you are building outbound automation, how much of your build time actually goes into the AI step?
Back to feed
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started