Agents & Agentic AI
Agentic AI
AI designed to plan, act, use tools, and adapt across multiple steps rather than only answer once.
Example
A research assistant plans several searches, evaluates results, and changes course when evidence is weak.
Why it matters
People use agentic AI for tasks with several steps. A tool can compare flights and book your choice.
You'll hear it
“It’s agentic AI. It can look up the order, issue the refund, and email the customer.”
Can you control it?
Sometimes
Sometimes. Some tools let you choose which steps run automatically, which tools or data they can use, and when they must ask before acting. Others set this for you.
Common questions
- What is the difference between agentic AI and an AI agent?
- Agentic AI names the approach. An AI agent is a particular system built to work that way.
- Is agentic AI the same as generative AI?
- No. Generative AI creates things such as writing or pictures; agentic systems may use that ability while carrying out tasks.
- Can agentic AI make mistakes?
- Yes. One wrong step can affect later steps, even when those later actions run as planned.