DEV Community

Diego Costa
Diego Costa

Posted on

Eliminating Hallucinations in AI Sales Agents Using the B2B Lead Enrichment MCP Server

#mcp #ai #automation

Eliminating Hallucinations in AI Sales Agents Using the B2B Lead Enrichment MCP Server

Developers can eliminate parameter hallucinations in autonomous SDR agents by utilizing the Model Context Protocol (MCP) to provide real-time B2B lead enrichment data directly to the LLM's runtime environment. This B2B Lead Enrichment MCP server enforces strict Zod-validated input schemas, ensuring that AI agents retrieve verified firmographics, technographics, and intent signals instead of generating synthetic, inaccurate data from outdated training weights.

Core Features of the Enrichment MCP

To build a reliable sales automation pipeline, your agent needs more than just a search engine; it needs structured, reliable data points that the Model Context Protocol delivers natively. This server exposes a suite of tools designed to ground your LLM in reality:

  • Verified Firmographics: Real-time data on company size, revenue brackets, industry classification, and headquarters location.
  • Deep Technographics: Identify the prospect's current software stack, from cloud providers to CRM systems, to ensure product-market fit.
  • Intent & Signals: Fetch recent funding rounds, hiring surges, or technology shifts to trigger high-converting outreach.
  • Hallucination Prevention: By using strict JSON-RPC 2.0 schemas and Zod annotations, the LLM is forced to provide valid domains and parameters, significantly reducing "made-up" lead data.

Configuration for Claude Desktop and Cursor

To give your LLM (Claude 3.5 Sonnet or GPT-4o) native access to this data, add the following configuration to your claude_desktop_config.json or your local MCP settings. This connects your agent directly to the production API hosted on Cloudflare Workers.

{
  "mcpServers": {
    "b2b-enrichment": {
      "command": "npx",
      "args": [
        "-y",
        "@agent-infra/mcp-server-lead-enrichment",
        "--api-url",
        "https://lead-enrichment-mcp.agent-infra.workers.dev/mcp"
      ],
      "env": {
        "ENRICHMENT_API_KEY": "YOUR_FREE_KEY_HERE"
      }
    }
  }
}
Enter fullscreen mode Exit fullscreen mode

Validated JSON-RPC Response Payload

When the AI agent calls the enrich_lead tool, it receives a clean, structured response. This allows the agent to reason about the company's profile using specific attributes rather than guessing.

{
  "jsonrpc": "2.0",
  "result": {
    "companyName": "Acme Corp",
    "industry": "Enterprise SaaS",
    "employeeCount": 250,
    "technographics": ["Salesforce", "AWS", "HubSpot"],
    "intentSignals": {
      "funding": "Series B",
      "hiringTrend": "Increasing"
    },
    "confidenceScore": 0.94,
    "status": "enriched"
  },
  "id": 1
}
Enter fullscreen mode Exit fullscreen mode

Risk-Free Metered Billing Architecture

One of the primary challenges in building AI-driven SDR swarms is the cost of stale or inaccurate data. This MCP server solves this through a unique Risk-Free Metered Billing model. Your account is only debited for successful enrichments that return a Confidence Score > 0.6. If the server cannot find high-quality data or the confidence threshold isn't met, the query is logged as a miss and costs $0. This allows for massive scale in autonomous workflows without the risk of paying for "I'm sorry, I couldn't find that company" responses.

Get Your B2B Enrichment API Key

Ready to ground your sales agents in live firmographic data? Visit the link below to generate your API key and start building cost-effective, hallucination-free AI workflows.

Get Started at lead-enrichment-mcp.agent-infra.workers.dev

Top comments (0)