Lessons from MWC: Most Companies Still Don’t Understand MCP скачать в хорошем качестве

Lessons from MWC: Most Companies Still Don’t Understand MCP

Last week we attended several AI events and meetups around MWC Barcelona, speaking with founders, builders, investors, and product teams about what’s actually happening in the AI ecosystem. One theme kept coming up: most companies still don’t really understand MCP. Instead of exposing their products to AI ecosystems, many teams are focused on adding AI agents inside their existing apps. That may improve usability, but it may also mean missing a much bigger shift. In this conversation we break down what we observed across the week, including: • Why companies are hesitant to adopt MCP • The difference between putting AI into your product vs putting your product into AI • Why startups are moving faster than incumbents • How tools like Figma, ChatGPT, and MCP servers are already changing product workflows • Why composability could reshape how software products compete We also share practical examples from our own workflows and conversations with builders during the week. If you’re building products, startups, or AI tools, this shift is happening faster than most companies realize. Chapters 00:00 Intro 01:00 What we learned at MWC Barcelona 02:00 Why most companies are adding AI inside their product 04:30 Why MCP is being overlooked 06:30 Startups vs incumbents in AI adoption 10:00 How companies are actually using AI today 11:30 Figma’s approach to MCP and AI workflows 17:30 How JC’s design workflow is changing with AI 22:30 The hidden impact of AI across industries 24:00 Why MCP needs better real-world examples 27:00 Composability: the real power of MCP 31:00 Why companies resist opening their data 33:00 Final thoughts