How I Find 2-5 Customers Daily Using Claude MCP + LinkedIn

Most founders spend hours scrolling LinkedIn looking for potential customers. The process is manual, inconsistent, and hard to scale. But if you connect Claude to real-time Google search data via MCP, you can systematically find 2-5 qualified prospects every day without leaving your terminal.

This post walks through a repeatable workflow that combines Claude's reasoning with live SERP data to identify, qualify, and prioritize LinkedIn prospects -- all in under 30 minutes per day.

The Core Idea

LinkedIn search is limited. You can filter by title and company, but you cannot cross-reference against what someone has recently published, what tools their company uses, or whether they are actively hiring for roles that signal buying intent. Google indexes all of that. By searching Google for LinkedIn profiles with specific contextual signals, you surface prospects that LinkedIn's own search misses.

With Scavio connected to Claude via MCP, you can run these searches conversationally and let Claude interpret the results, extract contact context, and draft personalized outreach -- all in one session.

Setting Up the MCP Connection

Connect Scavio to Claude Code with one command:

claude mcp add --transport http scavio https://mcp.scavio.dev/mcp \
  --header "x-api-key: YOUR_SCAVIO_API_KEY"

Once connected, Claude has access to the search_google tool. This is the only tool you need for this workflow.

Step 1: Define Your Buyer Signal Queries

The key is crafting Google queries that combine LinkedIn profile pages with buying signals. Here are patterns that work well:

site:linkedin.com/in "head of engineering" "series A" "hiring"
site:linkedin.com/in "VP of Marketing" "looking for" "analytics"
site:linkedin.com/in "CTO" "we just launched" 2026
site:linkedin.com/in "founder" "building" "API" "integration"

Each query targets a specific role combined with a contextual signal that indicates they might need your product. Ask Claude to run 3-5 of these queries per session using the search_google tool.

Step 2: Let Claude Qualify the Results

After each search, Claude receives structured SERP data including titles, snippets, and URLs. You can prompt Claude to:

• Extract the person's name, title, and company from the result
• Flag results where the snippet mentions a pain point you solve
• Cross-reference by searching for the company to understand their stack
• Rank prospects by how closely they match your ideal customer profile

A simple prompt like this works well:

Search Google for: site:linkedin.com/in "head of data" "e-commerce" 2026

For each result, extract the person's name, title, company, and any
buying signals from the snippet. Then search for each company to
determine what tools they currently use. Rank them by fit for a
real-time search API product.

Step 3: Generate Personalized Outreach Context

Once Claude has identified and ranked your prospects, ask it to generate outreach context for your top picks. Claude can search for recent articles the prospect wrote, conference talks they gave, or company announcements -- all using search_google.

This gives you personalized talking points that go beyond the generic connection request. You are referencing something they actually did or said, which dramatically increases response rates.

Making It a Daily Habit

The entire workflow takes 20-30 minutes:

• 5 minutes: Run 3-5 targeted searches
• 5 minutes: Review Claude's qualification and ranking
• 10 minutes: Deep-dive on top 2-5 prospects
• 5 minutes: Send personalized connection requests

Because Scavio returns structured JSON from Google results, the data is consistent and parseable every time. No scraping, no browser automation, no broken selectors. Just a search API call that returns clean data Claude can reason about.

The compounding effect is significant. At 2-5 prospects per day, you build a pipeline of 60-150 qualified leads per month -- all sourced from real-time search data rather than stale databases.