AI Dictionary of Terms

Agent

An autonomous AI entity that combines a language model with the ability to perceive its environment, reason about goals, plan actions, use tools, and execute multi-step tasks with minimal human intervention — the fundamental building block of agentic AI systems.

The Simple Version

If an LLM is a very smart brain in a jar, an Agent is that brain given a body, hands, and a to-do list. It can look at its surroundings (perceive), think about what to do (reason), grab tools to accomplish tasks (act), and check whether it succeeded (evaluate).

A chatbot waits for you to ask a question and gives one answer. An Agent is given a goal (“Book me a flight to London under $1,000”) and figures out all the steps on its own: search flights, compare prices, check your calendar, book the best option, and add it to your itinerary.

Detailed Explanation

An Agent is the concrete instantiation of agentic AI principles. While “Agentic AI” describes the paradigm, an “Agent” is the actual system doing the work.

Core Components of an Agent:

  1. Brain (LLM): The reasoning engine that processes information and makes decisions
  2. Memory:
    • Short-term: Current task context and conversation history
    • Long-term: Persistent knowledge about the user, past tasks, and learned patterns
  3. Tools: External capabilities the agent can invoke (search, APIs, databases, code execution)
  4. Planning Module: Ability to break goals into subtasks and sequence actions
  5. Action Loop: Perceive → Think → Act → Observe → Repeat

Agent Architectures:

ReAct (Reasoning + Acting): The agent alternates between reasoning about what to do and taking actions.

Thought: I need to find the current stock price of AAPL
Action: search_web("AAPL stock price today")
Observation: AAPL is trading at $195.50
Thought: Now I can answer the user's question
Action: respond("AAPL is currently at $195.50")

Plan-and-Execute: The agent creates a complete plan upfront, then executes each step.

Plan: 
1. Search for flights to London
2. Filter by price < $1000
3. Check user's calendar for availability
4. Book the best option
5. Add to calendar
Execute step 1... Execute step 2... etc.

Multi-Agent Systems: Multiple specialized agents collaborate on complex tasks.

Popular Agent Frameworks:

Key Characteristics

Business Context

Agents represent the next evolution of enterprise AI, moving from “AI as a tool” to “AI as a worker”:

Enterprise Applications:

Strategic Considerations:

Agent vs. Chatbot: | Aspect | Chatbot | Agent | |——–|———|——-| | Interaction | Single turn or conversation | Multi-step task execution | | Control | User drives every step | Agent drives toward goal | | Tools | Limited or none | Multiple external tools | | Memory | Conversation context | Task + long-term memory | | Autonomy | Low | High |

Real-World Analogy

A personal assistant vs. a search engine. A search engine (chatbot) answers your questions. A personal assistant (agent) takes your goal (“Plan my vacation”) and handles everything: researches destinations, checks your budget, books flights and hotels, creates an itinerary, and sets reminders. You just approve the final plan.

Code Example

# Simple agent using OpenAI's Agents SDK (conceptual)
from agents import Agent, Runner, function_tool

# Define tools the agent can use
@function_tool
def search_flights(destination: str, max_price: int) -> str:
    """Search for flights under a given price."""
    # In reality, this would call a flight API
    return f"Found 3 flights to {destination} under ${max_price}"

@function_tool
def book_flight(flight_id: str) -> str:
    """Book a specific flight."""
    return f"Flight {flight_id} booked successfully"

# Create the agent
travel_agent = Agent(
    name="Travel Assistant",
    instructions="You help users book flights. Always confirm details before booking.",
    tools=[search_flights, book_flight]
)

# Run the agent with a goal
result = Runner.run_sync(
    travel_agent,
    "Book me a flight to Tokyo under $1500 for next month"
)

print("Agent completed task:", result.final_output)
# The agent will:
# 1. Search for flights to Tokyo under $1500
# 2. Present options to user
# 3. Wait for confirmation
# 4. Book the selected flight
# 5. Return confirmation

Common Misconceptions

Sources & Further Reading