Agents & Agentic AI
Agent loop
The repeated cycle of observing, reasoning, acting, and checking results.
Example
An agent observes the current state, chooses an action, reviews the result, and repeats.
Why it matters
It helps an agent adjust its next step using what happened in the previous one.
You'll hear it
“It keeps trying the same failed action instead of changing approach.”
Can you control it?
Set by the team
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Can a loop get stuck?
- Yes. An agent may repeat actions without making progress if it misreads the result.
- Does every cycle have to take an external action?
- No. A cycle may decide to wait, ask for help or stop.
- Can limiting the number of cycles be useful?
- Yes. A limit can prevent a confused process from consuming time or resources indefinitely.