Agents

AI agents are systems that can operate autonomously or semi-autonomously, enabling them to perceive their surroundings, make informed decisions, and take actions to achieve specific objectives. Utilizing advanced artificial intelligence techniques—including machine learning, natural language processing, and reasoning—these agents interact intelligently with users, data, or other systems.

AI agents can range from straightforward implementations, such as chatbots that respond to inquiries, to more intricate systems, like self-driving cars that navigate through traffic. They typically function continuously, adapting and enhancing their performance over time by learning from new data.

Compass provides the following agent frameworks that you can leverage to build powerful agents to meet your specific needs.

• Autogen Framework: A library or system that helps create multi-agent workflows using large language models (LLMs) and enables LLM-powered agents to talk to each other and collaborate on tasks with minimal human intervention.
• LangChain Agent Framework: It is a part of the LangChain ecosystem that helps build AI agents capable of reasoning and taking actions using tools (like search engines, calculators, or databases) in a smart, iterative way.
• CrewAI: An open-source Python framework designed to facilitate the creation and orchestration of collaborative, role-based AI agents. These agents work together to tackle complex, multi-step tasks, much like a team of specialists collaborating on a project.​