← back to home
August 2025

Building Goatforce

Goatforce is a CRM dashboard powered by MCP (Model Context Protocol) servers, designed to be the one-stop shop for sales reps to manage their deals efficiently. The core idea was to eliminate the friction between different tools and create a unified interface that understands natural language commands.

The challenge was building a lightweight interface that could handle complex sales workflows while maintaining speed and reliability. We used MCP servers to bridge the gap between natural language processing and CRM operations, allowing sales reps to simply type what they want to accomplish rather than navigating through multiple screens and forms.

Why MCP Servers

Goatforce was built entirely leveraging MCP servers for seamless integration with Salesforce APIs. The MCP architecture allows for real-time data synchronization, automated deal tracking, and intelligent workflow management. By using MCP servers, we could maintain context across complex multi-step sales processes while ensuring data consistency and reliability.

The MCP server's strength lies in handling complex action chains without losing track of the sales pipeline. For example, when a sales rep types "update this deal to closed-won and schedule follow-up with the client," the system can execute multiple operations while adapting to changing CRM states and maintaining proper data relationships.

We tuned the compute allocation based on task complexity - giving minimal processing power for simple data updates and more computational resources for complex multi-step workflows. This approach balanced speed with reliability, ensuring that simple operations remained fast while complex processes had sufficient resources to complete successfully.

Websocket Challenge

// MCP Server Integration Example
const mcpClient = new MCPClient();
// Initialize connection to Salesforce
await mcpClient.connect('salesforce');
// Process natural language command
const result = await mcpClient.processCommand(
'Update deal ABC123 to closed-won',
{
complexity: 'medium',
context: 'sales-pipeline'
}
);