I’ve been experimenting with AI integration in software applications for some time, not as an isolated chatbot feature, but as part of the application architecture itself.
One of my first practical experiments was an AI-assisted chess database. The idea was simple: instead of navigating the application only through menus and forms, the user could interact with AI assistants able to understand the database, search games, add commentary, configure parts of the application and even play chess.
That project was an important starting point for me because it showed that AI could become another way of operating an application, not just something added on top of it.
Since then, the idea has evolved considerably. I’m now building a more general platform where AI, databases and application components share a common architecture. The application manages the context, exposes functions and resources to the models, and keeps control of the workflow.
These two videos show that evolution quite well:
[The chess database demo](https://youtu.be/BY7uQUw80ok)
[The current architecture](https://youtu.be/q-o2HpDuZ3E)
The first is where the idea started. The second is where it is heading now.