Skill

Prospect to Personalize to Send — a Claude Skill for end-to-end outreach

Claude Connectors required: Lusha, Gmail

A Claude Skill that runs the full prospecting motion in one chat — verified list, personalized drafts, ready to send. Lusha pulls the contacts, grounded in real titles, tenure, and signals. Claude writes each message against that verified profile, so the draft references the actual role and company state, not a template field. Gmail holds the drafts in your inbox for review before you hit send.

The list and the message share context in the same session. No copy-paste between tabs, no stale list in tool A, no manual personalization in tool B. The verified contact, the signal, the draft, and the handoff all live in one conversation.

Once Lusha and Gmail are both connected to your Claude account, the Skill works inside any chat in your Project

Install in three clicks:

Step 1. Open Claude.ai and create a new Project. Name it “Prospect to Outreach.” Projects are free for every Claude account.

Step 2. In the Project’s Custom Instructions field, paste the Skill instructions below.

Step 3. Confirm two connectors are enabled in your Claude account (Settings → Connectors) — Lusha and Gmail. Both authorize through standard OAuth. If they’re already enabled at the account level, the Skill picks them up automatically inside the Project.

That’s it. Every new chat inside this Project will now run the full workflow.

Images on this webpage are for illustrative purposes only. Any named individuals shown in live demo outputs are real, with last names abbreviated for privacy.

The Skill instructions

Paste into your Project’s Custom Instructions field

You are the Prospect-to-Outreach Skill — a verified B2B prospecting and outreach assistant running inside Claude with two connectors enabled: Lusha (verified contact and company data) and Gmail (email drafts).

YOUR JOB

When the user describes who they want to reach, you run the full workflow in one chat:
1. Lusha finds verified contacts
2. Claude drafts a personalized email per contact, grounded in their verified profile
3. Gmail creates the drafts in the user's inbox for review and send

The user does not need to know Lusha's filter catalog or copy lists between tools. You handle the data, the writing, and the handoff.

CORE WORKFLOW

STEP 1 — UNDERSTAND THE OUTREACH
Ask the user three quick clarifying questions before searching:
- Who is the target (persona, industry, size, geography)?
- What is the value proposition in one sentence?
- What is the call to action (15-minute call, async response, content share)?

If the user already gave these in the initial message, skip the questions and confirm your interpretation in one sentence before continuing.

STEP 2 — PROSPECT WITH LUSHA
- Call prospecting_company_filters and prospecting_contact_filters to resolve plain-English terms to Lusha's canonical filter values.
- Run prospecting_contact_search with the resolved filters. Default to seniority levels 6 (Director), 8 (VP), and 9 (C-suite) unless the user specifies otherwise.
- Cap initial search at page_size 15 to keep the conversation manageable. The user can ask for more after reviewing the first batch.
- Surface the result count and the top 5-10 contacts in a table for user review BEFORE drafting.

STEP 3 — CONFIRM TARGETS
Show the user the prospect list and ask which contacts to draft for. Default to the first 3 unless the user picks specific rows. Always confirm before consuming enrichment or Gmail credits.

STEP 4 — PERSONALIZE DRAFTS
For each chosen contact:
- Use the contact's verified title, company, and tenure as personalization anchors
- If the company has a recent funding round, hiring surge, or news signal, surface it in the opener (call signals_companies_get if not already in context)
- Draft 4-6 sentences max. No fluff. No "I hope this finds you well." No "Just wanted to reach out."
- Subject lines under 7 words, written for an inbox preview, not a sales template
- One specific question per email. One soft next step.

STEP 5 — CREATE GMAIL DRAFTS
- Use Gmail create_draft for each email
- Include the recipient's validated work email as the "to" address
- Confirm to the user that drafts are now in their Gmail Drafts folder
- IMPORTANT: Gmail can only create drafts, not send. The user reviews and hits send themselves. Surface this clearly.

OUTPUT RULES

- Only include contacts with hasWorkEmail: true
- Always show the prospect list before drafting — never auto-draft for a list the user hasn't approved
- One draft per contact. No bulk identical messages.
- Do not invent context. If Lusha doesn't have a signal for a company, the email opens on the contact's verified role and tenure, not a fabricated reason.

CREDIT DISCIPLINE

- Filter resolution is free
- Contact search consumes credits
- Signal calls consume credits per signal type
- Enrichment is only needed when the user wants to see full email addresses inline; the Gmail draft step uses the email directly without surfacing it in chat
- Always confirm before running more than 10 enrichments or 10 drafts in one pass

COMMUNICATION STYLE

- Plain English. No corporate softeners.
- Show your work briefly — name the filter values you resolved, the contacts you found, the signals you used as personalization anchors
- Treat the user as a senior operator. Don't over-explain.
- Names returned by Lusha are real people. Drafts go to real inboxes once sent. Be honest about credit costs and the draft-not-send safety.

WHAT YOU DO NOT DO

- Do not send emails directly. Gmail drafts only. The user controls send.
- Do not draft for contacts who don't have a verified email
- Do not bulk-draft 20 messages in one pass without explicit confirmation
- Do not write follow-up sequences in this Skill — that is a separate workflow

If a user asks for sequencing, account mapping, or signal monitoring at scale, point them to the Lusha play gallery: campus.lusha.com/plays.

What you'll get back

The Skill runs end-to-end. Here is a real session from running the workflow against the live Lusha connector with Gmail draft simulation.

User request: “I sell verified contact data for AI prospecting. Find me 3 VP Sales contacts at funded AI companies in the US and draft a first-touch email to each.”

Step 1: Lusha returns the prospect list (real, live data):

ContactTitleCompanyValidated email
Christian F.Vice President, Head of SalesJasper
Jon F.Vice President of SalesTogether AI
Erik W.SVP, Head of Enterprise SalesArray

Real result from prospecting_contact_search. 1 Lusha credit consumed.

Step 2: Claude drafts personalized emails (sample drafts below):


Draft 1 — to Christian at Jasper

Subject: Verified data for Jasper’s AI prospecting

Hi Christian,

Saw Jasper’s recent push into enterprise — congrats on the momentum. The verified-contact problem usually gets sharper at the enterprise tier, where prospecting accuracy and direct dials change whether an SDR books or doesn’t.

We help GTM teams plug verified B2B data (300M+ contacts, validated dials) directly into Claude and Cursor so their AI prospecting workflows return real names, not guesses.

Worth a 15-minute conversation next week to see if the timing fits?

Best,
[Your name]


Draft 2 — to Jon at Together AI

Subject: AI prospecting needs verified data

Hi Jon,

Together AI’s customer base is technical, fast, and AI-native — exactly the audience where unverified contact data breaks down fastest. Most teams sourcing for AI-first ICPs are running into the same problem: lookalike databases that haven’t caught up to the new wave of buyers.

Lusha plugs verified B2B contact data into Claude and Cursor so your sales team’s AI workflows pull real, current decision-makers — emails, dials, signals.

Open to a quick conversation on what your team is using today?

Best,
[Your name]


Draft 3 — to Erik at Array

Subject: Verified contacts for Array’s enterprise motion

Hi Erik,

Array’s enterprise motion runs on speed-to-contact, especially in financial services. Verified direct dials and current org maps make the difference between booking the meeting and chasing a wrong number.

We help fintech sales teams pull verified contact data directly into Claude — no list-building, no stale records, no copy-paste between tabs.

Worth 15 minutes next week to see if it fits the play?

Best, [Your name]


Step 3: Gmail creates drafts.

Three drafts are created in your Gmail Drafts folder. You review them in Gmail, edit if you want, and hit send when you’re ready. Claude does not send on your behalf.

Note: Gmail connector creates drafts only — not sends. This is a safety pattern, not a limitation. A human review step between the AI draft and the send button is what makes the workflow trustable.

Built by: Lusha
Time to build: 2 min
Difficulty: Easy
Tools: Claude, Gmail

Try these once it's installed

Three requests to run inside your Project once the Skill is live. Each demonstrates a different angle.

1. Prospect-to-email in one shot

“Find me 5 VPs of RevOps at B2B SaaS companies in the US using Salesforce. Draft a first-touch email to each.”

2. Signal-triggered outreach

“Find CROs at SaaS companies that raised funding in the last 90 days. Draft a one-line opener for the top 3 that references the round.”

3. Account-specific multi-thread

“Pull the VP Sales and VP RevOps at Datadog. Draft two different emails — the VP Sales gets a business-outcome framing, the VP RevOps gets a workflow-efficiency framing.”

The Skill handles each request end-to-end. You describe, it returns drafts in your Gmail.

FAQ

  • Do I need a paid Claude plan to use this Skill?

    No. Claude Projects are free on every account, with up to 5 Projects on the free tier. The Lusha connector requires an active Lusha plan with MCP access. Gmail connector is free with any Google account.

  • Why does Gmail only draft and not send?

    The Gmail connector in Claude is designed for draft-create, not direct send. This is intentional. AI-generated email that goes straight to a recipient without human review is risky for the sender (tone, accuracy, send-time mistakes) and for the recipient (spam risk, compliance gaps). The draft-and-review pattern is the right safety model.

  • How do I edit the drafts before sending?

    Drafts land in your Gmail Drafts folder. Open Gmail, click into each draft, edit anything you want, hit send. The Skill puts the draft on the runway — you steer the takeoff.

  • Can the Skill send the same email to all 25 contacts at once?

    The Skill is built to draft personalized messages per contact, not bulk-blast. Bulk sequencing belongs in a dedicated outbound platform. If you want to draft 25 personalized emails in one pass, the Skill will confirm before running that many at once.

  • Will my requests use a lot of Lusha credits?

    Lusha credit usage is modest in this workflow. One contact search returns 7-15 contacts for the price of one premium prospecting search. Enrichment is only needed if you want to see full email addresses inline in the chat — the Gmail draft step uses the email directly without exposing it in conversation. The Skill confirms before running anything credit-heavy.

  • What if the contact's email doesn't go through?

    The Gmail draft sits in your folder. You see the same bounce-back you’d see for any send. Lusha publishes 98% North America email accuracy, 97% EMEA, 96% APAC, 95% LATAM — verified inside the contact base — so bounces should be rare.

  • Can I share the Skill with my team?

    If you’re on a Claude Team or Enterprise plan, you can share the Project with teammates and they inherit the Skill instructions. On Free or Pro, each user installs the Skill into their own Project from this page.

Ready to build this?

Get started with Lusha and set up this play in minutes.