The way we engage with apps is undergoing a radical transformation, much like how the App Store reshaped the internet back in 2008. Instead of relying on traditional app icons on a phone, OpenAI is embedding apps directly into ChatGPT’s conversational interface. Now, developers of all sizes can seamlessly integrate their apps to be suggested contextually in chat or by direct mention. Early adopters like Adobe and Target bring unique functionalities—Adobe offers in-chat image editing controls, while Target enables shopping within conversations. This integration aims to shift the focus from app names to what users want to accomplish, enhancing multitasking through a more intuitive flow.
Behind the scenes, a new framework called the Model Context Protocol (MCP) supports these connections by allowing apps to share data, tools, and memory with AI in a streamlined fashion. Despite the excitement, challenges remain with app discovery and invocation within chats; users currently need to manually summon apps with specific commands, and seamless suggestions are still a work in progress. Furthermore, platform owners face the complex task of fairly recommending competing apps with similar features. Nevertheless, this emerging ecosystem promises to unlock creative new workflows and user experiences far beyond what traditional app design offers today.
