AI Agents

LLM-powered planning decomposes goals into executable steps. Tool use enables agents to interact with external systems (APIs, browsers, code execution). Memory systems maintain context across long task sequences. Evaluation frameworks measure agent reliability and task completion rates. Orchestration systems manage multi-agent collaboration.

AI coding agents complete development tasks from natural language specifications. Customer support agents resolve tickets end-to-end without human intervention. Research agents gather, synthesize, and report on topics autonomously. Sales agents automate prospecting, outreach, and qualification. QA agents test software automatically.

AI agents represent the frontier of generative AI where models transition from tools that respond to prompts to systems that independently accomplish multi-step goals. With 6 funders. The category has attracted billions in investment based on the thesis that AI agents could automate significant portions of knowledge work.