RESULTS, WRITTEN FROM THE LOGS
What actually happened when AI ran the outbound: the campaign numbers, the parts that broke, and the systems behind them. Everything here comes from our own pipeline and client campaigns we ran ourselves.
521 conversations in 12 days: exactly how the AI did it
A day-by-day breakdown of the recruitment campaign that started 521 LinkedIn conversations and booked 6 meetings in 12 days, including the two things that went wrong.
Read it →AI lead scoring in practice: cutting a 1,892-prospect list before it burns your account
How we score prospect lists with AI before any outreach happens, why we cut a third of most lists, and what happened to the campaign where the list was too big.
The enrichment waterfall: how AI finds the email address behind a warm LinkedIn lead
Our three-step waterfall for turning interested LinkedIn prospects into email conversations, with the real hit rates from 357 enriched contacts and the step most teams skip.
What 3,215 LinkedIn conversations taught us about acceptance and reply rates
Niche beats volume, boring sectors outperform crowded ones, and the first message decides everything. Field notes from 5,020 messages across our own and client campaigns.
Approval-first AI: 43 booked meetings and not one message sent without a human
The architecture that lets AI run your outbound without ever speaking for you unsupervised: default-reject queues, hard stops, and why the approval loop makes the AI better, not slower.
From Sales Navigator search to working prospect sheet in one evening
The exact pipeline that turns a raw LinkedIn search into a scored, enriched, campaign-ready prospect list without a VA, with the checks that stop garbage getting through.
Want the tools these articles describe? 10 free Claude skills →