Find Sales Prospects Using Google Reviews with Claude Code

Google Reviews are full of signals. A restaurant with complaints about slow service might need a POS system. A salon with “couldn’t book online” reviews needs a scheduling tool. A contractor with great work but terrible communication could use a CRM. Claude Code connected to AgentPatch can find these prospects, read their reviews, and build a targeted list.

Why This Matters

Cold outreach works better when you know the prospect’s actual problem. Generic “Hi, we help businesses like yours” emails get deleted. An email that says “I noticed several of your customers mentioned difficulty booking online” gets read.

The challenge is that mining reviews by hand is tedious. You’d need to search Google Maps, open each listing, read through reviews, and take notes. That’s an hour of work for maybe five good prospects. An agent can do the same thing in minutes.

Setup

Install the AgentPatch CLI (zero dependencies, Python 3.10+):

pip install agentpatch

Set your API key:

export AGENTPATCH_API_KEY=your_api_key

Then use it:

ap search "web search"
ap run agentpatch google-search --input '{"query": "test"}'

Get your API key from the AgentPatch dashboard.

Run this command to add AgentPatch as an MCP server:

claude mcp add -s user --transport http agentpatch https://agentpatch.ai/mcp \
  --header "Authorization: Bearer YOUR_API_KEY"

Replace YOUR_API_KEY with your actual key from the AgentPatch dashboard. Claude Code discovers all AgentPatch tools automatically.

The Workflow

Step 1: Find Businesses in a Category

Tell Claude Code what kind of businesses you’re targeting:

“Search Google Maps for hair salons in Portland, Oregon. Get me 10 results with their names, ratings, and review counts.”

Claude Code calls the Maps tool and returns a structured list.

Step 2: Analyze Reviews for Patterns

Now have the agent dig into the reviews:

“For each salon, search Google for their reviews. Look for recurring complaints: trouble booking, long wait times, rude staff, pricing issues, anything that comes up more than once. Summarize the top complaint pattern for each.”

Claude Code runs Google Search for reviews of each business and identifies the patterns. This is where the agent adds real value — it can read dozens of review snippets and pull out the common thread.

Step 3: Build a Prospect List

“Based on that research, which salons have a recurring problem that a booking or scheduling tool could solve? Rank them by how strong the signal is and give me a one-line summary for each.”

Prospect list — salons with booking/scheduling complaints:

1. Bloom Hair Studio (4.1★, 89 reviews)
   Signal: 12 reviews mention difficulty booking, 3 mention being turned away as walk-ins.
   Strongest signal in the set.

2. The Cutting Room (3.8★, 134 reviews)
   Signal: 8 reviews mention long wait times, several say "called but no answer."
   Likely no online booking system.

3. Velvet Salon (4.3★, 67 reviews)
   Signal: 5 reviews mention wanting to book online, one mentions a broken website link.
   Good candidate — high ratings but friction in booking.

4. Mane Street Hair (3.5★, 42 reviews)
   Signal: Mixed complaints. Some about booking, but also quality and pricing issues.
   Weaker fit — problems may be deeper than scheduling.

Remaining 6 salons: No strong booking-related complaint patterns found.

You now have 3 strong prospects with specific talking points pulled directly from their customers’ words.

Wrapping Up

Review mining turns cold outreach into warm outreach. You’re not guessing what the prospect needs — their customers already told you. Claude Code with AgentPatch handles the research: Maps for discovery, Search for review analysis. Start at agentpatch.ai.

← Back to blog