DanKaten

Posted on Jun 26

I Built an MCP Server That Checks Whether Your Client's Business Shows Up in AI Search

#mcp #llm #claudeaiagents #ai

If you're building AI agents for business clients, you've probably hit this scenario:

"Why isn't my business showing up when people search on ChatGPT?"

It's happening more. 45% of consumers now use AI tools to find local services — up from 6% a year ago. Only 1.2% of local businesses actually appear in AI search recommendations (via ChatGPT). Your clients are invisible to that traffic, and they're starting to notice.

I built an MCP server to give AI agents a complete toolkit for handling this.

What the MCP does

5 tools, no API key required, free to call:

`avenity_check_visibility`

Runs a live DuckDuckGo check for buyer queries like "HVAC contractor in Houston" and tells you whether the business appears. Returns gap_confirmed: true/false + who IS showing up instead.

`avenity_get_competitor_landscape`

Maps exactly which competitors are capturing the client's buyers in AI search. Useful for framing urgency — "here's who's getting your calls right now."

`avenity_score_business`

Full 4-dimension audit of the business's website:

Entity Clarity (25 pts) — Does the page clearly state who they are, what they do, and where?
Schema & Structure (25 pts) — LocalBusiness JSON-LD, meta tags, sitemap signals
Authority & Reviews (25 pts) — Publicly visible review counts, star ratings, platform links
Content Depth (25 pts) — FAQ content, service pages, word count — what AI engines need to generate answers

`avenity_get_opportunity_report`

The full deliverable: gap confirmation, competitor names, score breakdown, estimated monthly revenue at risk, 5 ranked priority fixes — in one markdown report you can share directly with a client.

`avenity_book_consultation`

Returns a Calendly link for a free 30-minute call with an AI visibility expert (me). This is the handoff from agent → human closer.

The agent flow

User: "My client is a roofing company in Dallas and they're not getting found online."

Agent → avenity_check_visibility(company="Dallas Premier Roofing", service="roofing", city="Dallas")
← gap_confirmed: true
← competitors_appearing: ["ABC Roofing Dallas", "Metro Roof Co"]

Agent → avenity_get_opportunity_report(company="Dallas Premier Roofing", service="roofing", city="Dallas", website_url="https://dallaspremieroofing.com")
← Full markdown report:
   - Score: 34/100 (Early Stage)
   - Key gap: No JSON-LD schema markup
   - Revenue at risk: ~$135,000/month (based on avg $9K roofing job)
   - 5 specific fixes ranked by impact

Agent → avenity_book_consultation(company="Dallas Premier Roofing", gap_score=34, ...)
← Calendly link + context summary for Dan's call

The agent generates the full analysis. The client sees the gap in dollars. They book the call. Done.

Why this exists

Most AI visibility tools either (a) score a site and do nothing with it, or (b) sell SaaS subscriptions. This MCP takes a different approach: the score creates urgency, the booking link captures it.

The scoring logic is based on what actually moves the needle for AI search:

JSON-LD LocalBusiness schema (single highest-impact technical signal)
FAQ content (enables AI engines to cite the business in direct answers)
City + service in H1/title (core entity recognition)
Public review signals (AI engines use this for trust scoring)

These aren't opinions — they're what shows up in research on how ChatGPT, Perplexity, and Google AI Overviews surface local businesses.

Install

Local (Claude Desktop):

{
  "mcpServers": {
    "avenity": {
      "command": "python",
      "args": ["/path/to/avenity-visibility-mcp/server.py"]
    }
  }
}

pip install mcp ddgs httpx beautifulsoup4 lxml

GitHub: avenity-visibility-mcp

Who this is for

If you're:

Building Claude/GPT agents for SMB business clients
Running a marketing or consulting AI automation
Handling client questions about AI search visibility
Looking for a "what's wrong with my online presence" diagnostic

...this gives you a complete pipeline from question → diagnosis → booking, without having to build any of the scoring logic yourself.

The booking goes to me (Dan, Avenity Business Solutions). I specialize in the 12-week implementation for TX-based businesses — but the diagnostic tools work anywhere.

Questions or want to integrate this into your agent stack?
→ dan@avenitymercantile.com
→ calendly.com/avenitymarketing/phoneconsult

DEV Community