Agents are AI Subunits that receive special system instructions and can be given access to expert knowledge data via RAG Retrieval Augmented Generation or Dynamic Endpoint Injection for real-time data.
There are generally three types of Agents:
- Expert Agent
- Function Calling Agent
- Research Agent
Most of the Agents however, are hybrid agents and combine these capabilities.
As an alternative to entering a plain query into the chat box, you can use an Agent. Some of the agents are pinned to the chat page. Others can be called by either typing "@name of agent" in the chat input filed or by clicking the "View all" button and then selecting the agent from the directory. You can pin and arrange your agents by clicking the "Edit" button.
You can search for Agents by keyword or scrolling and then simply clicking on the "Chat Now" button on the desired character card.
Alternatively you can activate an Agent directly from the Chat Box by typing the @(NAME OF AGENT) symbol:
The use of AI Agents can be beneficial for several reasons:
-
Consistency: Ensuring that the AI's responses adhere to a particular character or role allows for a more coherent and predictable dialogue, which is especially useful in storytelling, gaming environments, or role-playing scenarios.
-
Specialization: By assigning a character with certain expertise, the AI can tailor responses to be more specialized and relevant to that field. For example, a doctor persona might be created to provide more authoritative responses in medical scenarios.
-
User Experience: Agents can make interactions more engaging and relatable, emulating social and conversational norms that users might expect when interacting with another human being rather than a machine.
-
Context Management: Agents can help manage the context of a conversation better by maintaining a specific viewpoint or knowledge base, avoiding confusion that could arise from more general or broad responses.
-
Training Focus: When creating AI applications, personas can streamline the process of training and fine-tuning by focusing on the specific language, technical language, or attitudes that the persona should exhibit.
-
Expressiveness: Agents allow for a wider range of emotional and stylistic expression, which can be more captivating and memorable for users.
-
Education and Training: In pedagogical applications, personas can simulate scenarios with various stakeholders, providing a more immersive and practical learning experience.
-
Diverse Perspectives: Using multiple Agents can introduce a variety of perspectives and voices, enriching the discourse and promoting critical thinking.
-
Accountability and Scope Definition: Clear definitions of Agents help in setting boundaries for AI capabilities, making it easier to manage user expectations and establish clear guidelines for AI behavior.
-
Research and Testing: Agents can be used to run controlled experiments to observe how different types of AI Agents interact with users and which are more effective in certain situations.
-
Brand Alignment: Brands can use specific Agents to align with their image and values, ensuring that AI interactions represent them accurately.
-
Entertainment and Leisure: Creatively, Agents can contribute to games, interactive fiction, and other entertainment forms, providing a more immersive and engaging experience.
Using Agents within Nionium AI can be viewed as a framework for guiding generative AI towards a certain style, tone, or knowledge domain, which is designed to enhance the user's interaction and satisfaction with the system.