What is an AI agent, and how is it different from a single LLM call?

beginner

Re-explain

An AI agent is an LLM given a goal, a set of tools, and permission to take several steps on its own. A single LLM call is one turn: text in, text out, then it is over.

	Single LLM call	Agent
Steps	Exactly one	Keeps going until the goal is met
Who picks the next action	Your code	The model, at runtime
Tools	None, or one round you handle	Called repeatedly, results fed back in
State	Stateless	Carries context across steps
Cost and latency	Predictable	Varies run to run

The real difference is control. You trade a predictable path for the ability to handle tasks whose steps are unknown up front. That is why every agent needs step limits, spend budgets, and logs.

Rewriting in plainer words…

This answer doesn't lend itself to a diagram - it reads best . No credits were charged.

Why there's no diagram: “”

The interactive diagram is below the answer - jump to diagram ↓ · Below it, the related concept . Jump to it ↓

The diagram below the answer is the concept . Jump to it ↓

Tailored explanation · switch back to · ·

What should the new diagram focus on?

Guided practice for AI Agents One question at a time. Answer out loud, get graded, see what you missed.

How well did you know this?

AI:

Saved in this browser - sign in to keep your review list.

How should your speech become text?

Listening… your words appear above as you speak - tap Stop when you're done.

Recording · cr - tap Stop & transcribe when you're done.

Transcribing with AI…

Voice:

Keep going - a few more words and AI can grade it.

Interview lens How interviewers actually use this question

Likely follow-ups

Walk me through what happens in one turn of the agent loop. The model reasons and picks a tool call, your code runs it, the result goes back into context, and the loop ends on a final answer or a limit.
When the model 'calls a tool', what actually runs it, and what does the model get back? The model only emits a tool name and arguments. Your app executes it and returns the result as a message. The model never runs code itself.
When would you build a fixed workflow instead of an agent? When the steps are known up front. A path your code picks is cheaper, faster and easier to test. Use an agent only where you can't script the path.
How do you stop an agent from looping forever? A hard step cap and spend budget, a check for repeated identical tool calls, and per-step logs so you can see why it looped.
How is evaluating an agent different from evaluating a single LLM call? Score the whole trajectory and the end result, not one output. Runs vary, so test many runs and check tool choices and step counts.

What you can say

An AI agent is an LLM with a goal, a set of tools, and permission to take several steps on its own.
A single LLM call is one turn. Text goes in, text comes out, and it's over.
In an agent, the model picks the next action at runtime, calls tools, reads the results and keeps going until the goal is met.
It carries context from step to step, while a single call is stateless and your code decides what happens next.
So the real difference is control. My code no longer picks the path, the model does.
That lets you handle tasks whose steps you can't know up front, but cost and latency vary run to run, so every agent needs step limits, spend budgets and logs.

Weak answers to avoid

Treats one function call as an agent Your code handling one round of tool calls is still a single call. It becomes an agent when the model picks the next action repeatedly until the goal is met.
Defines an agent as a chatbot with memory Memory alone doesn't make an agent. The defining change is who controls the next step. Say the model chooses actions and calls tools in a loop.
Pitches agents as strictly better Agents give up predictable cost and latency. Say when a single call is enough, and name step limits, spend budgets and logs as the price of using one.
Says 'autonomous AI' with no mechanism Buzzwords don't show understanding. Describe the loop instead. The model picks a tool, your code runs it, the result goes back in, and it repeats.

Interview lens

Likely follow-ups, what you can say, and the weak answers to avoid.

Want a quick review of the fundamentals? See the AI Agents cheatsheet.

← Back to all AI Agents questions

Pro · $10/mo

47 of 54 AI Agents answers are in Pro.

Full answers, code samples, and AI explanations that go simpler or deeper. Cancel anytime.

Full answers + code
AI explanations, simpler or deeper
1,000 AI credits / month
Cancel anytime

Get Pro → See plans & pricing →