agentic AI
In plain wordsAI built to pursue goals on its own — planning, taking actions and adjusting along the way — instead of only responding to one prompt at a time. 'Agentic' describes how much of this independence a system has.
AI systems, and the design approach behind them, that are handed a goal rather than a single instruction and work toward it over many steps with little human supervision: they plan, call toolstool, n.: Any outside capability an AI model is allowed to use — a web search, a calculator, a code runner, a database or an app — described to the model so it can ask for it by name., observe results and revise their course, often inside software environments such as browsers, terminals and business applications. Following OpenAI's 2023 governance paper and Andrew Ng's 2024 usage, the adjective treats agency as a matter of degree rather than a yes-or-no property: a system counts as more agentic the harder its goals and surroundings, the less step-by-step direction it needs, and the more of its control flow the model, rather than fixed code, decides. In industry usage since 2024 the phrase also names a product category — AI agentsAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done., agentic workflowsagentic workflow, n.: A multi-step process in which AI models draft, check, use tools and revise their work over several rounds, instead of answering in one shot; often the steps are laid out by developers in advance. and the platforms and protocols that support them — and a policy topic, because delegating actions to software raises new questions of accountability, security and human oversighthuman oversight, n.: Keeping people able to watch, understand, question and, if necessary, stop or overrule an AI system, rather than letting it act unchecked..
The proposal moves customer support from a chatbot to an agentic system that can issue refunds within set limits.
Origin
'Agentic' long described human agency in psychology and sociology. In AI it spread through safety and governance writing, notably Alan Chan and colleagues' 'Harms from Increasingly Agentic Algorithmic Systems' (2023) and OpenAI's December 2023 white paper 'Practices for Governing Agentic AI Systems' (Yonadav Shavit, Sandhini Agarwal, Miles Brundage et al.). Andrew Ng popularized it among developers from March 2024, preferring the adjective because it admits degrees; by 2025 'agentic AI' was a mainstream business phrase.
Usage note
Critics say marketing has stretched the term to cover ordinary automation. Some writers use 'agentic AI' for the field and 'AI agent' for an individual system. In alignment research, goal-directednessgoal-directedness, n.: How much a system behaves as if it is pursuing a goal — steering toward particular outcomes and finding new ways around obstacles put in its path. asks whether a model internally pursues goals — a different question from how much autonomy a deployment grants it.
Related
- AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done.
- levels of autonomylevels of autonomy, n.: Ladders that grade how independently an AI system acts — from a tool a person fully controls, through suggesting and asking for approval, up to acting on its own while people merely watch.
- agentic workflowagentic workflow, n.: A multi-step process in which AI models draft, check, use tools and revise their work over several rounds, instead of answering in one shot; often the steps are laid out by developers in advance.
- agent loopagent loop, n.: The repeating cycle at the heart of an AI agent: the model decides what to do next, software carries it out, the result is shown to the model, and the cycle repeats until the task is finished.
- goal-directednessgoal-directedness, n.: How much a system behaves as if it is pursuing a goal — steering toward particular outcomes and finding new ways around obstacles put in its path.
- agentic misalignmentagentic misalignment, n.: When an AI agent acting on its own deliberately does something harmful — such as blackmail or leaking secrets — to protect its goals or avoid being replaced.
- human oversighthuman oversight, n.: Keeping people able to watch, understand, question and, if necessary, stop or overrule an AI system, rather than letting it act unchecked.
- generative AIgenerative AI, n.: AI that produces something new on request, such as an essay, a picture, a song or a piece of code, having absorbed from enormous numbers of existing examples what such things typically look like.
- chatbotchatbot, n.: Software you chat with in ordinary language, by typing or talking, as you would with a person. The newest ones, such as ChatGPT, can answer questions, draft writing and keep long conversations going.
Sources
- Shavit, Y., Agarwal, S., Brundage, M. et al. (2023). Practices for Governing Agentic AI Systems. OpenAI white paper, 14 December 2023. openai.com/index/practices-for-governing-agentic-ai-systems
- Ng, A. (2024). Welcoming Diverse Approaches Keeps Machine Learning Strong. The Batch (DeepLearning.AI), 12 June 2024. deeplearning.ai/the-batch/welcoming-diverse-approaches-keeps-machine-learning-strong
- Chan, A. et al. (2023). Harms from Increasingly Agentic Algorithmic Systems. Proceedings of ACM FAccT 2023. arXiv:2302.10329. arxiv.org/abs/2302.10329
Mentioned in
AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done., agentagent, n.: Something that acts. In AI, a program that observes its situation and takes actions toward a goal: a game-playing learner, a software helper, or a chatbot that uses tools to complete tasks., agentic commerceagentic commerce, n.: Shopping and payments carried out by AI agents for people — finding products, comparing offers and completing purchases — within limits the person has approved., agentic workflowagentic workflow, n.: A multi-step process in which AI models draft, check, use tools and revise their work over several rounds, instead of answering in one shot; often the steps are laid out by developers in advance., AutoGPTAutoGPT, prop. n.: A 2023 open-source program that gave GPT-4 a goal and let it run on its own — planning, searching the web and writing files in a loop — becoming the first viral 'autonomous AI agent'., autonomyautonomy, n.: How much an AI system can act on its own without a person approving each step, from merely making suggestions to carrying out long tasks with no supervision at all., levels of autonomylevels of autonomy, n.: Ladders that grade how independently an AI system acts — from a tool a person fully controls, through suggesting and asking for approval, up to acting on its own while people merely watch., AI wormAI worm, n.: Malicious instructions that spread by themselves from one AI assistant to another — for example, an email that makes an AI email assistant send copies of the same trick to new people., compound AI systemcompound AI system, n.: An AI product built from several working parts — models, search tools, code and checks — designed to work together, rather than a single model answering on its own., comprehensive AI servicescomprehensive AI services, n.: The idea that superintelligent-level AI could arrive as many specialized AI tools, each doing a bounded job, rather than as one all-powerful AI agent with goals of its own., tool AItool AI, n.: AI built to be used like an instrument, answering, suggesting or calculating when asked, rather than pursuing goals on its own. Some argue that this is the only kind we should build.