Your cold email infrastructure is only as good as the list you send to. You can have perfectly warmed domains, verified DNS records, and a sending platform dialed to the right volume — and still wreck your deliverability in a single campaign if the list is bad. Invalid emails bounce. Bounces spike your sender score. A 5% bounce rate is enough to trigger spam filters across every mailbox you spent weeks warming up.
The list is the single highest-leverage piece of the entire cold email system. Get it right and every email you send has a real person on the other end. Claude builds the full list — sourcing, validation, deduplication, enrichment, and scoring — so what you load into your sending platform is clean from the start.
How Claude Helps You Build This
Instead of manually pulling contacts from multiple tools, cross-referencing in spreadsheets, and hoping your emails are valid, Claude runs the entire list-building pipeline. It starts from your ICP definition, sources contacts, validates every email address, removes duplicates, enriches with real buying signals, and scores each prospect so you know who to email first.
To do this, Claude uses three skills:
- company-contact-finder — locates decision-maker contacts at target companies with verified email addresses and LinkedIn profiles.
- lead-qualification — filters and scores every contact against your ICP criteria, ranking multi-signal leads highest.
- contact-cache — checks all leads against previously contacted prospects and deduplicates across sourcing channels.
You provide a one-paragraph description of your ideal customer. Claude builds the list.
The Workflow
Step 1: Define your ICP as filtering criteria
Give Claude a one-paragraph description of who you're targeting — role, company size, industry, tools they use, and what problem you solve for them. Claude expands this into structured filtering criteria:
- Job titles and seniority levels to target
- Company size ranges (headcount and revenue)
- Industries and sub-industries
- Technology signals (what tools they already use)
- Geographic filters if relevant
These criteria become the search parameters for every sourcing channel.
Step 2: Source contacts across multiple channels
Using the company-contact-finder skill, Claude pulls prospects from Clay, Apollo, or LinkedIn Sales Navigator based on your filtering criteria. It cross-references multiple databases — contacts that appear in only one tool get caught when you query two or three.
For a typical B2B campaign, Claude sources 500–2,000 contacts per run depending on how narrow your ICP is. Each contact comes with name, title, company, email, and LinkedIn URL where available.
Step 3: Validate every email address
This is the step most people skip — and the one that burns infrastructure fastest. Claude runs every email through validation to catch:
- Invalid or non-existent addresses
- Catch-all domains (risky but usable with caution)
- Role-based addresses (info@, sales@) that hurt deliverability
- Disposable email providers
Email validation costs roughly $1 per 1,000 addresses. On a list of 1,500 contacts, that's $1.50 to protect weeks of warmup work. Claude flags everything above a safe bounce threshold and removes it before the list goes anywhere near your sending platform.
Step 4: Deduplicate across sources and existing contacts
If you're sourcing from multiple tools or running multiple campaigns, duplicates are inevitable. Using the contact-cache skill, Claude cross-references your new list against:
- Contacts from other sourcing runs
- Your existing CRM contacts
- Any suppression lists (customers, partners, competitors)
Double-emailing the same prospect from different campaigns is one of the fastest ways to get marked as spam. Claude catches it before it happens.
Step 5: Enrich with buying signals
A validated email isn't enough — you need context for personalization and prioritization. Claude enriches each contact with real signals:
- Recent LinkedIn activity and post topics
- Job changes in the last 90 days
- Company funding rounds or growth indicators
- Technology stack changes
- Hiring patterns that signal budget allocation
These signals do double duty — they feed your personalization when you write sequences, and they feed the scoring model in the next step.
Step 6: Score and prioritize
Using the lead-qualification skill, Claude scores every contact based on ICP fit and intent signals. The output is a ranked list with three tiers:
- Tier 1 — strong ICP match + active buying signals. Email these first.
- Tier 2 — good ICP match, no active signals. Solid second batch.
- Tier 3 — partial match. Hold or nurture.
You review the Tier 1 list, confirm the scoring makes sense, and approve it for campaign loading. One cold email practitioner reported that validating and cleaning their list alone dropped their bounce rate from 8% to under 1% — the difference between domains that survive and domains that get blacklisted.
What You Walk Away With
After running this workflow, you have:
- Sourced contacts — 500–2,000 prospects pulled from multiple databases, matched to your ICP
- Validated emails — every address verified, invalid and risky contacts removed
- Deduplicated list — no double-sends across campaigns or existing contacts
- Enriched profiles — LinkedIn activity, job changes, funding signals, tech stack data attached to each contact
- Scored and tiered prospects — ranked by ICP fit and buying intent, ready to load
Why This Matters
A clean list protects the infrastructure you built and puts real prospects on the other end of every email. Bad lists don't just waste sends — they actively destroy your sender reputation. Bounce rates above 5% trigger spam filters. Spam filters tank deliverability across every mailbox and domain you set up.
This workflow gives you a list you can trust. Once it's loaded, you're ready for the final step: launching and scaling your campaigns without blowing past the volume limits that keep you in inboxes.
Build your prospect list with Gooseworks