384 rows verified, 116 marked, none invented

Why does everyone end up running a waterfall?

Coverage is uneven and uncorrelated. Provider A is strong on US tech and weak on European industrials; provider B is the reverse. Stopping at the first confident hit lifts match rates over any single source. Clay built a product around the pattern, and Clearbit (folded into HubSpot as Breeze) and Apollo remain common single-source picks, as of this writing.

Wiring that sequence, feeding it and settling the rows where two providers disagree is the job. A companion does it for all 500 rows, in the provider accounts you already pay for.

Running the waterfall in your own accounts

Hunter, Lusha, Snov and Apollo are native integrations, so the sequence runs as tool calls against accounts you already pay for: try Hunter, fall through to Lusha, fall through again, record which source answered.

For the rows the databases miss, the companion opens the company site, reads the team page and checks the person still works there, on every leftover row rather than the first twenty. Output lands in the CSV or staged into your CRM with each write pausing. See how to automate lead enrichment.

What are honest numbers for a real batch?

A 500-row conference list: Hunter resolves 296 emails, Lusha adds 61 of the remainder, Apollo another 27. Session reading corrects 38 job changes where the title was right and the company was not.

Final: 384 rows verified and sourced, 116 marked unfound. Your time: ten minutes writing the instruction, fifteen spot-checking forty rows.

Point it at the right list first

Enrichment multiplies whatever you give it, so get the list right first. Treat "found" and "safe to send" as two claims with verification between them. Personal-contact data carries jurisdiction-specific rules, which is a call for whoever owns your privacy policy to make once and write into the instruction.

Frequently asked questions