A warm LinkedIn reply is worth money, but LinkedIn is a terrible place to close. Threads get buried, notifications get missed, and the moment a prospect changes jobs the conversation is gone. Email is where deals actually progress. Which creates the unglamorous middle step of all outbound: you have a name and a LinkedIn profile, and you need a working email address. Here is the exact waterfall our system runs, with real numbers from our own pipeline, where it has found and verified 357 addresses so far.
The waterfall, in order
The system tries three sources in sequence and stops at the first hit. Order matters, because the sources differ in cost, accuracy and coverage.
- LinkedIn-native lookup first. A provider that resolves emails from the LinkedIn profile itself. Strong on UK and EU coverage, and it returns the address the person actually uses professionally rather than a guessed pattern.
- Contact-database enrichment second. Name plus company into a B2B enrichment API. The useful trick here is that good providers return the person's current professional email even when their LinkedIn profile lags a job move, which happens more than you would think.
- The company website last. The published contact address: info@, hello@, office@. Nobody is proud of this step, which is exactly why most teams skip it and lose the lead entirely. A well-written note to an office inbox that opens with who you were talking to and why gets forwarded more often than cold email folklore suggests. It is the last resort, and it works often enough to keep.
The real numbers, not the vendor deck
Across our current pipeline, the distribution looks like this: the majority of interested leads either already had an address on file or got one from the first lookup, a meaningful tail came from the second-step enrichment API, and roughly one in six of the hard cases ended up using the website route. No single provider gets you everything. The waterfall exists because coverage is a patchwork, and any vendor claiming a 95% hit rate is measuring against an easy list.
Two operational rules make the waterfall compound:
- Every found address is written back to the database immediately. A lead is never enriched twice. Over months this quietly becomes an owned asset: hundreds of verified, warm-sourced addresses that no data vendor can sell you.
- Failures stay in the queue. If all three steps miss, the lead is not discarded. It is logged and retried later, because people update profiles, companies publish team pages, and databases refresh.
What the AI adds beyond plumbing
Chaining three APIs is a Tuesday afternoon of engineering. The part that needs intelligence is everything around it.
The AI decides who is worth enriching at all, because enrichment costs money per lookup and an unscored list burns budget on prospects who would never buy. It validates what comes back, because a pattern-guessed address that bounces damages the sending domain that your whole email channel depends on. And it writes the first email around the LinkedIn conversation that earned the address, so the prospect reads a continuation, not a cold start. The address is the plumbing. The context is the sale.
The mistake that costs teams the most
Waiting. The half-life of a warm reply is measured in days. Teams that batch their enrichment "for Friday" email prospects whose interest has already cooled. Our orchestrator enriches within hours of a lead turning interested, drafts the email the same run, and has it waiting for approval the same morning. Speed between channels is a conversion lever most teams never pull, and it costs nothing.
The whole pipeline, from interested reply to enriched address to drafted email in your voice, is what AI Sales runs every morning. Or start smaller: the free ICP Scorer tells you who is worth enriching in the first place.