Home Think Topics AI Agents What are AI agents?
Explore IBM's AI agent solution Subscribe for AI updates
Illustration showing a customer interacting with a customer service chatbot

Published: 3 July 2024
Contributor: Anna Gutowska

What are AI agents?

An artificial intelligence (AI) agent refers to a system or program that is capable of autonomously performing tasks on behalf of a user or another system by designing its workflow and utilizing available tools.

AI agents can encompass a wide range of functionalities beyond natural language processing including decision-making, problem-solving, interacting with external environments and executing actions. These agents can be deployed in various applications to solve complex tasks in various enterprise contexts from software design and IT automation to code-generation tools and conversational assistants. They leverage the advanced natural language processing techniques of Large Language Models (LLMs) to comprehend and respond to user inputs step-by-step and determine when to call on external tools.

How to choose the right AI foundation model

Use this model selection framework to choose the most appropriate model while balancing your performance requirements with cost, risks and deployment needs.

How AI agents work

At the core of AI agents are large language models (LLMs). For this reason, AI agents are often referred to as LLM agents. Traditional LLMs, such as IBM® Granite™ models, produce their responses based on the data used to train them and are bounded by knowledge and reasoning limitations. In contrast, agentic technology utilizes tool calling on the backend to obtain up-to-date information, optimize workflow and create subtasks autonomously to achieve complex goals. In this process, the autonomous agent learns to adapt to user expectations over time. The agent's ability to store past interactions in memory and plan future actions encourages a personalized experience and comprehensive responses.1 This tool calling can be achieved without human intervention and broadens the possibilities for real-world applications of these AI systems. The approach that AI agents take in achieving goals set by users is comprised of these three stages:

Goal initialization and planning

Although AI agents are autonomous in their decision-making processes, they require goals and environments defined by humans.2 There are three main influences on autonomous agent behavior: 

  • The team of developers that design and train the agentic AI system. 
  • The team that deploys the agent and provides the user with access to it.
  • The user that provides the AI agent with specific goals to accomplish and establishes available tools to use.

Given the user's goals and the agent’s available tools, the AI agent then performs task decomposition to improve performance.3 Essentially, the agent creates a plan of specific tasks and subtasks to accomplish the complex goal.

For simple tasks, planning is not a necessary step. Instead, an agent can iteratively reflect on its responses and improve them without planning its next steps.

Reasoning using available tools

AI agents base their actions on the information they perceive. Often, AI agents do not have the full knowledge base needed for tackling all subtasks within a complex goal. To remedy this, AI agents use their available tools. These tools can include external datasets, web searches, APIs and even other agents. After the missing information is retrieved from these tools, the agent can update its knowledge base. This means that each step of the way, the agent reassesses its plan of action and self-corrects.  

To help illustrate this process, imagine a user planning their vacation. The user tasks an AI agent with predicting which week in the next year would likely have the best weather for their surfing trip in Greece. Since the LLM model at the core of the agent does not specialize in weather patterns, the agent gathers information from an external database comprised of daily weather reports for Greece over the past several years. Despite acquiring this new information, the agent still cannot determine the optimal weather conditions for surfing and so, the next subtask is created. For this subtask, the agent communicates with an external agent that specializes in surfing. Let’s say that in doing so, the agent learns that high tides and sunny weather with little to no rain provide the best surfing conditions. The agent can now combine the information it has learned from its tools to identify patterns. It can predict which week next year in Greece will likely have high tides, sunny weather and a low chance of rain. These findings are then presented to the user. This sharing of information between tools is what allows AI agents to be more general-purpose than traditional AI models.3

Learning and reflection

AI agents use feedback mechanisms, such as other AI agents and human-in-the-loop (HITL), to improve the accuracy of their responses. Let’s return to our previous surfing example to highlight this. After the agent forms its response to the user, the agent stores the learned information along with the user’s feedback to improve performance and adjust to user preferences for future goals. If other agents were used to reach the goal, their feedback may also be used. Multi-agent feedback can be especially useful in minimizing the time that human users spend providing direction. However, users can also provide feedback throughout the agent's actions and internal reasoning to better align the results with the intended goal.Feedback mechanisms improve the AI agent's reasoning and accuracy, which is commonly referred to as iterative refinement.3 To avoid repeating the same mistakes, AI agents can also store data about solutions to previous obstacles in a knowledge base.

Agentic versus non-agentic AI chatbots

AI chatbots use conversational AI techniques such as natural language processing (NLP) to understand user questions and automate responses to them. These chatbots are a modality whereas agency is a technological framework. 

Non-agentic AI chatbots are ones without available tools, memory and reasoning. They can only reach short-term goals and cannot plan ahead. As we know them, non-agentic chatbots require continuous user input to respond. They can produce responses to common prompts that most likely align with user expectations but perform poorly on questions unique to the user and their data. Since these chatbots do not hold memory, they cannot learn from their mistakes if their responses are unsatisfactory.

In contrast, agentic AI chatbots learn to adapt to user expectations over time, providing a more personalized experience and comprehensive responses. They can complete complex tasks by creating subtasks without human intervention and considering different plans. These plans can also be self-corrected and updated as needed. Agentic AI chatbots, unlike non-agentic ones, assess their tools and use their available resources to fill in information gaps. 

Reasoning paradigms

There is not one standard architecture for building AI agents. Several paradigms exist for solving multi-step problems. 

ReAct (Reasoning and Action) 

With this paradigm, we can instruct agents to first "think" and plan before deciding which tools to use and iteratively improving upon responses. These Think-Act-Observe loops are used to solve problems step by step. Through the prompt structure, agents can be instructed to reason slowly and to display each "thought".4 The agent's verbal reasoning gives insight into how responses are formulated. In this framework, agents continuously update their context with new reasoning. This can be interpreted as a form of Chain-of-Thought prompting

ReWOO (Reasoning WithOut Observation)

The ReWOO method, unlike ReAct, eliminates the dependence on tool outputs for action planning. Instead, agents plan upfront. Redundant tool usage is avoided by anticipating which tools to use upon receiving the initial prompt from the user.  This is desirable from a human-centered perspective since the user can confirm the plan before it is executed. The ReWOO workflow is made up of three modules. In the planning module, the agent anticipates its next steps given a user's prompt. The next stage entails collecting the outputs produced by calling these tools. Lastly, the agent pairs the initial plan with the tool outputs to formulate a response. This planning ahead can greatly reduce token usage and computational complexity as well as the repercussions of intermediate tool failure.

Types of AI agents

AI agents can be developed to have varying levels of capabilities. A simple agent may be preferred for straightforward goals to limit unnecessary computational complexity. In order of simplest to most advanced, there are 5 main agent types:

Simple reflex agents

Simple reflex agents are the simplest agent form that grounds actions on current perception. This agent does not hold any memory, nor does it interact with other agents if it is missing information. These agents function on a set of so-called reflexes, or rules. This means that the agent is preprogrammed to perform actions that correspond to certain conditions being met. If the agent encounters a situation that it is not prepared for, it cannot respond appropriately. The agents are only effective in environments that are fully observable granting access to all necessary information.7

Example: A thermostat that turns on the heating system at a set time every night. The condition-action rule here is, for instance, if it is 8 PM, then the heating is activated.

Model-based reflex agents

Model-based reflex agents use both their current perception and memory to maintain an internal model of the world. As the agent continues to receive new information, the model is updated. The agent’s actions depend on its model, reflexes, previous precepts and current state. These agents, unlike simple reflex agents, can store information in memory and can operate in environments that are partially observable and changing. However, they are still limited by their set of rules.7

Example: A robot vacuum cleaner. As it cleans a dirty room, it senses obstacles such as furniture and adjusts around them. The robot also stores a model of the areas it has already cleaned to not get stuck in a loop of repeated cleaning. 

Goal-based agents

Goal-based agents have an internal model of the world and also a goal or set of goals. These agents search for action sequences that reach their goal and plan these actions before acting on them. This search and planning improve their effectiveness when compared to simple and model-based reflex agents.8

Example: A navigation system that recommends the fastest route to your destination. The model considers various routes that reach your destination, or in other words, your goal. In this example, the agent’s condition-action rule states that if a quicker route is found, the agent recommends that one instead.

Utility-based agents

Utility-based agents select the sequence of actions that reach the goal and also maximize utility or reward. Utility is calculated using a utility function. This function assigns a utility value, a metric measuring the usefulness of an action or how “happy” it will make the agent, to each scenario based on a set of fixed criteria. The criteria can include factors such as progression toward the goal, time requirements, or computational complexity. The agent then selects the actions that maximize the expected utility. Hence, these agents are useful in cases where multiple scenarios achieve a desired goal and an optimal one must be selected.8

Example: A navigation system that recommends the route to your destination that optimizes fuel efficiency and minimizes the time spent in traffic and the cost of tolls. This agent measures utility through this set of criteria to select the most favorable route.

Learning agents

Learning agents hold the same capabilities as the other agent types but are unique in their ability to learn. New experiences are added to their initial knowledge base, which occurs autonomously. This learning enhances the agent’s ability to operate in unfamiliar environments. Learning agents may be utility or goal-based in their reasoning and are comprised of four main elements:8

  • Learning: This improves the agent’s knowledge by learning from the environment through its precepts and sensors.
  • Critic: This provides feedback to the agent on whether the quality of its responses meets the performance standard.
  • Performance: This element is responsible for selecting actions upon learning.
  • Problem generator: This creates various proposals for actions to be taken. 

Example: Personalized recommendations on e-commerce sites. These agents track user activity and preferences in their memory. This information is used to recommend certain products and services to the user. The cycle repeats each time new recommendations are made. The user’s activity is continuously stored for learning purposes. In doing so, the agent improves its accuracy over time. 

Use cases of AI agents
Customer experience

AI agents can be integrated into websites and apps to enhance the customer experience by serving as a virtual assistants, providing mental health support, simulating interviews and other related tasks.10 There are many no-code templates for user implementation, making the process of creating these AI agents even easier. 


AI agents can be used for various real-world healthcare applications. Multi-agent systems can be particularly useful for problem-solving in such settings. From treatment planning for patients in the emergency department to managing drug processes, these systems save the time and effort of medical professionals for more urgent tasks.11

Emergency response

In case of natural disasters, AI agents can use deep learning algorithms to retrieve the information of users on social media sites that need rescue. The locations of these users can be mapped to assist rescue services in saving more people in less time. Therefore, AI agents can greatly benefit human life in both mundane tasks and life-saving situations.12

Benefits of AI agents

Task automation

With the ongoing advancements in generative AI, there is a growing interest in workflow optimization using AI, or intelligent automation. AI agents are AI tools that can automate complex tasks that would otherwise require human resources. This translates to goals being reached inexpensively, rapidly and at scale. In turn, these advancements mean human agents do not need to provide direction to the AI assistant for creating and navigating its tasks. 

Greater performance

Multi-agent frameworks tend to outperform singular agents.13 This is because the more plans of action are available to an agent, the more learning and reflection occur. An AI agent incorporating knowledge and feedback from other AI agents specializing in related areas can be useful for information synthesis. This backend collaboration of AI agents and the ability to fill information gaps are unique to agentic frameworks, making them a powerful tool and a meaningful advancement in artificial intelligence.

Quality of responses

AI agents provide responses that are more comprehensive, accurate and personalized to the user than traditional AI models. This is extremely important to us as users since higher-quality responses typically yield a better customer experience. As previously described, this is made possible through exchanging information with other agents, using external tools and updating their memory stream. These behaviors emerge on their own and are not preprogrammed.14

Risks and limitations

Multi-agent dependencies

Certain complex tasks require the knowledge of multiple AI agents. When implementing these multi-agent frameworks, there is a risk of malfunction. Multi-agent systems built on the same foundation models may experience shared pitfalls. Such weaknesses could cause a system-wide failure of all involved agents or expose vulnerability to adverse attacks.9 This highlights the importance of data governance in building foundation models and thorough training and testing processes.

Infinite feedback loops

The convenience of the hands-off reasoning for human users using AI agents also comes with its risks. Agents that are unable to create a comprehensive plan or reflect on their findings, may find themselves repeatedly calling the same tools, invoking infinite feedback loops. To avoid these redundancies, some level of real-time human monitoring may be used.9

Computational complexity

Building AI agents from scratch is both time-consuming and can also be very computationally expensive. The resources required for training a high-performance agent can be extensive. Additionally, depending on the complexity of the task, agents can take several days to complete tasks.14

Best practices

Activity logs 

To address the concerns of multi-agent dependencies, developers can provide users with access to a log of agent actions.6 The actions can include the use of external tools and describe the external agents utilized to reach the goal. This transparency grants users insight into the iterative decision-making process, provides the opportunity to discover errors and builds trust.


Preventing AI agents from running for overly long periods of time is recommended. Particularly, in cases of unintended infinite feedback loops, changes in access to certain tools, or malfunctioning due to design flaws. One way to accomplish this is by implementing interruptibility. Maintaining control of this involves allowing human users the option to gracefully interrupt a sequence of actions or the entire operation. Choosing if and when to interrupt an AI agent requires some thoughtfulness as some terminations can cause more harm than good. For instance, it may be safer to allow a faulty agent to continue assisting in a life-threatening emergency than to completely shut it down.5

Unique agent identifiers

To mitigate the risk of agentic systems being used for malicious use, unique identifiers can be used.6 If these identifiers were to be required for agents to access external systems, there would be greater ease in tracing the origin of the agent's developers, deployers and its user. This would be particularly helpful in case of any malicious use or unintended harm done by the agent. This level of accountability would provide a safer environment for these AI agents to operate.

Human supervision

To assist in the learning process for AI agents, especially in their early stages in a new environment, it can be helpful to provide occasional human feedback. This allows the AI agent to compare its performance to the expected standard and adjust accordingly. This form of feedback is helpful in improving the agent’s adaptability to user preferences.Apart from this, it is best practice to require human approval before an AI agent takes highly impactful actions. For instance, actions ranging from sending mass emails to financial trading should require human confirmation.8 Some level of human monitoring is recommended for such high-risk domains.

Related products and solutions
IBM watsonx Assistant™

Leverage IBM watsonx Assistant, an enterprise-grade AI chatbot, which is equipped with generative abilities to produce accurate, conversational answers. 

Explore watsonx Assistant Try watsonx Assistant

IBM watsonx Code Assistant™

Increase developer productivity through the use of IBM watsonx Code Assistant powered by IBM Granite foundation models. Features include code generation, matching and modernization.

Explore watsonx Code Assistant Try watsonx Code Assistant

IBM watsonx Orchestrate™

Lighten workload and improve productivity using generative AI assistants with IBM watsonx Orchestrate, powered by LLMs. 

Explore watsonx Orchestrate Try watsonx Orchestrate
Related resources Train a software agent to behave rationally with reinforcement learning

Use this machine learning technique to identify actions for states within an environment.

Enterprise generative AI virtual assistants: IBM® watsonx™

Build AI assistants that empower individuals to do work without expert knowledge.

AI Augmented Software Development with Agents (Assistants)

Agents, also known as assistants, are specialized versions of large language models (LLMs) that are pre-prompted and fine-tuned to support specific roles.

What is a virtual agent?

A virtual agent combines natural language processing, intelligent search and robotic process automation (RPA) in a conversational UI–typically a chatbot.

Take the next step

Train, validate, tune and deploy generative AI, foundation models and machine learning capabilities with IBM watsonx.ai™, a next-generation enterprise studio for AI builders. Build AI applications in a fraction of the time with a fraction of the data.

Explore watsonx.ai Book a live demo

1 Andrew Zhao, Daniel Huang, Quentin Xu, Matthieu Lin, Yong-Jin Liu, and Gao Huang, "Expel: Llm agents are experiential learners," Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 38, No. 17, pp. 19632-19642, 2024, https://ojs.aaai.org/index.php/AAAI/article/view/29936 (link resides outside ibm.com)

Yonadov Shavit, Sandhini Agarwal, Miles Brundage, Steven Adler, Cullen O’Keefe, Rosie Campbell, Teddy Lee, Pamela Mishkin, Tyna Eloundou, Alan Hickey, Katarina Slama, Lama Ahmad, Paul McMillan, Alex Beutel, Alexandre Passos and David G. Robinson, “Practices for Governing Agentic AI Systems,” OpenAI, 2023, https://arxiv.org/pdf/2401.13138v3 (link resides outside ibm.com).

3 Tula Masterman, Sandi Besen, Mason Sawtell, Alex Chao, “The Landscape of Emerging AI AgentArchitectures for Reasoning, Planning, and Tool Calling: A Survey,” arXiv preprint, 2024, https://arxiv.org/abs/2404.11584 (link resides outside ibm.com)

4 Gautier Dagan, Frank Keller, and Alex Lascarides, "Dynamic Planning with a LLM," arXiv preprint, 2023. https://arxiv.org/abs/2308.06391 (link resides outside ibm.com)

Binfeng Xu, Zhiyuan Peng, Bowen Lei, Subhabrata Mukherjee, Yuchen Liu, and Dongkuan Xu, "ReWOO: Decoupling Reasoning from Observations for Efficient Augmented Language Models," arXiv preprint, 2023, https://arxiv.org/abs/2305.18323 (link resides outside ibm.com)

Devjeet Roy, Xuchao Zhang, Rashi Bhave, Chetan Bansal, Pedro Las-Casas, Rodrigo Fonseca, and Saravan Rajmohan, "Exploring LLM-based Agents for Root Cause Analysis," arXiv preprint, 2024,  https://arxiv.org/abs/2403.04123 (link resides outside ibm.com)

Sebastian Schmid, Daniel Schraudner, and Andreas Harth, "Performance comparison of simple reflex agents using stigmergy with model-based agents in self-organizing transportation." IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion, pp. 93-98, 2021, https://ieeexplore.ieee.org/document/9599196 (link resides outside ibm.com)

8 Veselka Sasheva Petrova-Dimitrova, “Classifications of intelligence agents and their applications,” Fundamental Sciences and Applications, Vol. 28, No. 1, 2022.

9 Alan Chan, Carson Ezell, Max Kaufmann, Kevin Wei, Lewis Hammond, Herbie Bradley, Emma Bluemke, Nitarshan Rajkumar, David Krueger, Noam Kolt, Lennart Heim and Markus Anderljung, “Visibility into AI Agents,” The 2024 ACM Conference on Fairness, Accountability, and Transparency, pp. 958-973, 2024, https://arxiv.org/abs/2401.13138 (link resides outside ibm.com)

10 Lei Wang, Chen Ma, Xueyang Feng, Zeyu Zhang, Hao Yang, Jingsen Zhang, Zhiyuan Chen, Jiakai Tang, Xu Chen, Yankai Lin, Wayne Xin Zhao, Zhewei Wei, and Jirong Wen, “A survey on large language model based autonomous agents,” Frontiers of Computer Science, Vol. 18, No. 6, 2024, https://link.springer.com/article/10.1007/s11704-024-40231-1 (link resides outside ibm.com)

11 Jaya R. Haleema, Haleema, N. C. S. N. Narayana, “Enhancing a Traditional Health Care System of an Organization for Better Service with Agent Technology by Ensuring Confidentiality of Patients’ Medical Information,” Cybernetics and Information Technologies, Vol. 12, No. 3, pp.140-156, 2013, https://sciendo.com/article/10.2478/cait-2013-0031 (link resides outside ibm.com).

12 Jingwei Huang, Wael Khallouli, Ghaith Rabadi, Mamadou Seck, “Intelligent Agent for Hurricane Emergency Identification and Text Information Extraction from Streaming Social Media Big Data,” International Journal of Critical Infrastructures, Vol. 19, No. 2, pp. 124-139, 2023, https://arxiv.org/abs/2106.07114 (link resides outside ibm.com)

13 Junyou Li, Qin Zhang, Yangbin Yu, Qiang Fu, and Deheng Ye. "More agents is all you need." arXiv preprint, 2024, https://arxiv.org/abs/2402.05120 (link resides outside ibm.com)

14 Joon Sung Park, Joseph O'Brien, Carrie Jun Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein, "Generative agents: Interactive simulacra of human behavior," Proceedings of the 36th Annual ACM Symposium on User Interface software and Technology, pp. 1-22, 2023, https://dl.acm.org/doi/10.1145/3586183.3606763 (link resides outside ibm.com)