This article explains what agentic means and what an agentic system is in the context of AI and LLM system development.

---

What does agentic mean?

Agentic refers to the capacity to act independently and make choices. It implies a sense of agency, or the ability to exert control over one's own actions and decisions.

Here are some key aspects of agentic behavior:

• Autonomy: The ability to act independently and make choices without external influence.
• Initiation: The ability to take the first step and initiate actions.
• Persistence: The ability to continue pursuing goals despite setbacks or challenges.
• Responsibility: The willingness to take ownership of one's actions and their consequences.

What are Agentic LLM Systems?

In the context of LLM development, agentic refers to the ability of a language model to act independently and autonomously. This means that the model can:

Generate creative and original content: Without relying solely on the data it has been trained on. Adapt to new situations: By learning from new information and adjusting its responses accordingly. Initiate actions: Rather than simply responding to prompts, the model can take the initiative and suggest actions or ideas. Demonstrate understanding: By comprehending the nuances of language and context.

Essentially, an agentic LLM is one that goes beyond simple text generation and exhibits a higher level of intelligence and adaptability.

Langgraph is the main tool in 2024 for development of agentic systems. It is built on top of Langchain.

Reference: https://langchain-ai.github.io/langgraph/concepts/high_level/