ChatGPT Apps
By the end of 2025, OpenAI will open for ChatGPT app submissions.
With more than 800 million monthly users, this will create an entirely new market. Companies can tap into this opportunity by bringing their unique solutions to AI platforms. To succeed, these solutions must be accessible to large language models and designed to be useful in a chat-based environment.
The new SEO
How do I get my solution in front of users?
Traditionally, companies relied on SEO (search engine optimization) to increase their visibility online. This meant creating content with relevant keywords and earning backlinks, helping your website rank higher in search engines like Google.
However, over the last three years, user search behavior has been shifting toward AI chat platforms such as ChatGPT. If this trend continues, companies will need to optimize their solutions for AI discoverability to maintain visibility. But how do you make your website — or your service — easily discovered and usable by AI?
Optimizing for LLM Discovery & Use
We believe that within a few months, building an app for ChatGPT or other AI chat platforms will be as essential as having a website. Most AI chat systems currently rely on tools that search the web and attempt to read and extract information. This involves parsing page structures, waiting for content to load, and interpreting the result — a slow and unreliable process. Instead, we can eliminate the guesswork by explicitly providing structured information to the model. We can even deliver custom UI components to present that information to users. This should be front-of-mind for any solution that aims to be accessible to AI systems.
The current standard for making servers and solutions interoperable with AI agents is the Model Context Protocol (MCP). An MCP server defines and implements the tools an AI agent can use. Major AI platforms such as ChatGPT and Cursor are already adopting this protocol, enabling developers to extend their capabilities. This is the core of optimizing for AI now:
Build MCP servers and interfaces on top of them.
The next big thing?
Now that MCP tools can present data with user-friendly interfaces, we are entering a landscape full of opportunities for new and innovative solutions. But there is also an urgent need: existing solutions must be adapted to this format. If AI becomes the dominant interface, companies that fail to adapt risk losing customers to competitors who do.
