Agentic models

Autonomous AI systems that leverage Large Language Models (LLMs) for goal-directed planning, tool use (APIs, databases), and complex, multi-step task execution with minimal human intervention.

Agentic models represent the next evolution of Generative AI, moving past reactive chatbots to proactive, goal-driven systems. They operate via a core loop: perception, planning, action, and self-reflection. The LLM acts as the reasoning engine, orchestrating specialized components like memory systems (for context) and external tool access (APIs) for real-world interaction. Frameworks such as Microsoft's AutoGen and LangGraph enable multi-agent collaboration, allowing specialized agents (e.g., a 'Price Monitor' or 'Inventory Manager') to work together autonomously on complex workflows. Gartner projects one-third of all enterprise software applications will include agentic AI by 2028: this is the shift from digital assistants to virtual collaborators.

https://www.freecodecamp.org/news/agentic-ai-handbook/