Mar 11, 2024
The key features of the platform:
1. The ‘storyline’
Create your own flow or storyline in the configuration file. The AI will follow the storyline – you’re not just using prompting to guide the chatbot. This is a full-fledged state machine that keeps the bot on topic.
2. State machine + state updates
The chatbot analyzes what’ going on in the conversation and updates state accordingly, just like someone keeping the scoreboard updated while watching a football game, for example.
The state and the conversational history are stored in a persistent database.
Connecting two disparate systems is the secret to our success:
AI – conversation and reasoning
State machine and database – checkpoints and structure for the AI.
3. Transitions
When a user hits a checkpoint in the storyline, the chatbot is dynamically updated with a new prompt to advance the story/flow from that point.
4. Plugins
The code uses a plugin system so that additional elements may be added to the storyline easily.
The unique, proprietary highlights of the platform source code:
1. Enhancing long term memory
The state machine architecture along with persistent storage of state and chat history in a database lets the chatbot remember user details and past interactions to maximize a customized user experience, even in lengthy and complex conversations.
2. Dynamic AI-driven interactions
Meanwhile, dynamic state updates allow the chatbot to instantaneously adapt to new information going on in the chat – and respond the way you want it to, not the other way around. This vastly outperforms standard chatbots that operate on predefined response patterns.
3. Scalable and customizable design
With its modular design and abstracted logic, Project Dragon can be easily customized and scaled across various industries, offering a versatile platform that supports sophisticated chatbot applications far beyond the capabilities of traditional chatbot frameworks.