MCP

⚡ Key Points Claude provides the AI reasoning and natural language intelligence. Salesforce remains the single source of truth for business records, permissions, and automated workflows. MCP (Model Context Protocol) acts as the universal plug-and-play adapter that allows AI applications to securely talk to external systems. An MCP Server securely exposes specific, approved Salesforce data (like Opportunities or Cases) to the AI. Bottom Line: MCP does not bypass Salesforce security. It standardizes how AI safely accesses authorized APIs. 🗺️ Module Map CLAUDE FOUNDATION ├── What Is Claude? ├── Claude Model Families ├── Claude Reasoning ├── Claude Tool Use ├── AI Agents ├── MCP (Model Context Protocol) ├── MCP Architecture ├── MCP Security & Governance ├── Claude + MCP + Salesforce ← [You Are Here] └── Claudeforce Architecture 💬 In plain words: Think of this setup as a highly efficient corporate team. ...

💬 In plain words: Think about a new employee joining your company. First, security verifies their identity at the front door ( Authentication ). Next, their badge dictates which specific rooms they can enter ( Authorization ). Implementing the Model Context Protocol ( MCP ) works exactly the same way. When an AI agent like Claude tries to pull data from Salesforce, enterprise governance ensures it proves who it is acting on behalf of, and strictly limits what it is allowed to see or change. ⚡ 1-Minute Summary Enterprise AI demands intelligence paired with ironclad security. MCP bridges AI with external systems, but it relies on robust Authentication (verifying identity), Authorization (controlling actions), Permissions (defining data access), and Governance (monitoring and controlling usage) to keep your corporate data secure. 🗺️ Module Map: Claudeforce Architecture CLAUDE FOUNDATION ├── What Is Claude? ├── Claude Models ├── Claude Reasoning...

💬 In Plain Words: Imagine if every single AI application required a custom-built connector to talk to Salesforce, GitHub, local databases, or your file system. The Model Context Protocol (MCP) eliminates this nightmare. Promoted heavily by Anthropic, MCP is an open standard that acts like a universal "USB-C cable for AI." It allows AI models to seamlessly discover and connect to external data sources and tools using a single, unified protocol. 🗺️ Module Map CLAUDE FOUNDATION ├── What Is Claude? ├── Claude Models ├── Claude Reasoning ├── Claude Tool Use ├── AI Agents ├── Model Context Protocol (MCP) ← Current Topic ├── MCP Architecture └── Claudeforce Architecture 🎬 Real-Life Example: One AI Employee, Many Systems A sales manager tells their AI assistant: "Prepare me for today's customer meetings." To do this, the AI needs CRM data from Salesforce, proposal documents from Google Drive, support tickets from Jira, and calendar events from...

💬 In plain words: An action is a specific capability you give to an AI agent. In Salesforce Agentforce, there are five types of actions you can build. Choosing the right type for a specific business requirement is a core architectural skill, not just a matter of personal preference. ⚡ 1-Minute Summary: The five action types are: Flow, Apex, Prompt Template, External Service, and MCP . The AI model chooses which action to run by reading the action's description . Therefore, you must write descriptions as explicit selection criteria. If you use Apex, the action should return a decided value and a reasoning string, never just a raw data record. The Module Map: Agent Capabilities MODULE 4 Root: "What can the agent actually DO, and who decides when?" ├─ 4.1 The Five Action Types (You are here) ├─ 4.2 Descriptions as Selection Criteria ├─ 4.3 Return a Decision, Not Data └─ 4.4 Actions vs. Reasoning Actions ...