Agent
Microsoft distinguishes between agents and AI assistants.
Agents: Specialized systems tailored for specific tasks, described as the "new apps for an AI-powered world."
Assistants: Primarily support general activities such as drafting emails source.
Anthropic acknowledges various definitions of agents. Their blog post identifies two key types:
- Fully autonomous systems: Operate independently over extended periods.
- Prescriptive implementations: Follow predefined workflows.
Salesforce broadly defines AI agents as systems capable of understanding and responding to customer inquiries without human intervention. They categorize agents into six types, notably including:
- Simple reflex agents
- Utility-based agents
TechCrunch emphasizes the ambiguity surrounding the term "AI agent," noting the absence of a universally accepted definition source.
An "Agentic AI moment" refers to events where an AI unexpectedly demonstrates significant autonomy and resourcefulness, similar to the well-known "ChatGPT moment." These moments highlight AI capabilities that surpass user expectations in areas like competence, creativity, and problem-solving source.
AI Agents are programs where LLM outputs control the workflow.
Any system leveraging LLMs will integrate the LLM outputs into code. The influence of the LLM's input on the code workflow is the level of agency of LLMs in the system.
Note that with this definition, "agent" is not a discrete, 0 or 1 definition: instead, "agency" evolves on a continuous spectrum, as you give more or less power to the LLM on your workflow.
The table below illustrates how agency varies across systems:
