Every major AI company building on MCP yet hardly anyone outside of engineering can explain what it actually does.
In this video I break down the Model Context Protocol, the open source standard from Anthropic that is becoming the most important piece of infrastructure in the AI stack. I sat down with Marc Benioff, c-founder and CEO of Salesforce, right after he announced 30 new AI capabilities for Slackbot and talked about why Salesforce is betting on Slack becoming the operating system for AI-powered work.
We get into how MCP works, why Slack is now both an MCP server and an MCP client, and why Salesforce believes the future of AI belongs to the platforms that connect all of your AI investments together.
Timestamps
00:00 Marc Benioff on Slack Becoming the AI Interface
00:45 Why AI Protocols Matter More Than Models
01:04 What is MCP (Model Context Protocol) Explained
01:59 MCP Adoption: 97 Million Downloads in 18 Months
02:26 How Salesforce Built an AI Ecosystem Around Slack
02:58 Slack as MCP Server and MCP Client Explained
04:33 Salesforce Agentforce Stack Architecture Breakdown
05:29 Data 360 and the Federated Data Layer Challenge
06:38 The Race to Own the AI Agent Interface
07:37 Multi Sensory Models and Why Engineers Should Learn MCP