Crewline: a self-driving recruiting engine for a market that isn't on the job boards

Crewline builds self-driving compaction systems for construction sites, and hires the operators and robotics talent to run them. Both sides of that equation share one problem: the best people are rarely on a job board or an Apollo list. The operators are on jobsites, not updating their profiles. The perception and controls engineers are already employed and not looking. Crewline needed reach into markets where the standard tools come up empty, and it needed that reach to run without adding headcount.

The challenge

Crewline's targets live off the beaten path. Their customer ICP is US contractors that self-perform earthworks and site prep, owner-operators and superintendents who don't respond to generic email. Their hiring ICP is specialized robotics talent that no volume of inbound will surface. Early tests confirmed it: email-only outreach at low ICP fit produced a 0.31% reply rate. Reaching this market takes real sourcing, real personalisation, and more than one channel. Doing that by hand does not scale.

What we built

We deployed a self-driving outbound engine that sources talent and buyers across channels, engages them with personalised multichannel sequences, retargets non-responders automatically, and keeps one clean source of truth behind it all. Sourcing runs off-LinkedIn, rebuilding contact data from the open web and maps directories so the engine reaches contractors that job boards and LinkedIn leave untouched. Every prospect is enriched, deduplicated against a master database, scored for ICP fit, and phone-verified before it ever enters a sequence.

What we deployed: Apify and Claude Code plus a dedicated Google Maps scraper for off-LinkedIn sourcing, Clay for enrichment, deduplication, blocklist filtering and ICP scoring, Attio as the CRM and single source of truth, Lemlist for email plus LinkedIn plus call sequences, n8n for orchestration, and Slack alerts for live reply monitoring.

How it runs

Weekly, new companies flow from sourcing into a raw table, get enriched and ICP-scored in Clay, then promote into campaign tables that push straight to Lemlist and into Attio for the calling queue. Sequences engage across email, LinkedIn and calls, with non-responders retargeted automatically. Replies surface in Slack in real time, the client-side caller works a prioritised Attio call list, and every "not relevant" flag feeds back to tune the sourcing. It runs as a closed loop with a regular check-in, not a pile of one-off campaigns.

The outcome

The signal came through clearly: multichannel peer outreach carries this market where email alone cannot. The LinkedIn-led founder-to-founder motion became the account's strongest channel, with the best-performing sequence reaching a double-digit reply rate against a low-single-digit baseline for email-only. The same engine now runs both sales and robotics-recruiting campaigns off one system. Crewline gets predictable reach into a fragmented, off-the-grid market, and a pipeline that compounds week over week without new hires.