For businesses and tech enthusiasts seeking the next major breakthrough, 'agentic AI' is being heralded as the future. While the term is rapidly becoming a marketing staple across the industry, it represents a significant evolution from the generative AI chatbots currently dominating the landscape.
What Exactly is Agentic AI?
Unlike chatbots that primarily converse and generate content, agentic AI systems are designed to take direct action. A new report published on Tuesday, 18th November 2025, by researchers from the Massachusetts Institute of Technology and Boston Consulting Group details this shift. The report, which surveyed more than 2,000 business executives globally, describes agentic AI as a 'new class of systems' that can autonomously 'plan, act, and learn.'
These systems are not merely tools waiting for instructions; they are increasingly behaving like autonomous teammates, capable of executing multi-step processes and adapting in real-time. This distinction is crucial. While chatbots like the original ChatGPT, which debuted three years ago, rely on large language models to predict text, agentic AI is built for action.
The Industry's Vision for Autonomous Agents
Major tech players, including OpenAI, Amazon, Google, IBM, Microsoft, and Salesforce, are actively developing this technology. Swami Sivasubramanian, Vice President of Agentic AI at Amazon Web Services, explained the core difference. He stated that while a chatbot provides ideas, an agentic AI system takes the initiative to act upon them, breaking down a high-level goal into a series of steps.
Sivasubramanian, who assumed his new role earlier this year, believes agentic AI will be one of the most significant transformations since the advent of cloud computing. For consumers, the first practical applications are likely to emerge in areas like online shopping and travel, where an AI could make purchases or bookings based on a set budget and preferences.
The Future Potential and Expert Perspectives
Looking further ahead, experts envision AI agents handling more complex tasks. Thomas Dietterich, a professor emeritus at Oregon State University, expressed his desire for an agent that could manage medical bills or act as a 'personal shield' against email spam and phishing. He acknowledges the immense potential of AI systems that have the 'freedom and responsibility' to refine their goals, though he cautions against the overuse of the 'agentic' label for simple automated tasks.
The term itself has deeper roots. Milind Tambe, a professor at Harvard University who has researched multi-agent systems for three decades, finds the recent popularity of the adjective 'agentic' amusing. The concept of software agents acting autonomously has been discussed since the first International Conference on Multi-Agent Systems in 1995.
The push to popularise 'agentic' was also aided by prominent AI researcher Andrew Ng, co-founder of Coursera, over a year ago. He valued the term for its technical precision, suggesting it helped distinguish substantive discussions from mere marketing fluff.
As Google searches for 'agentic' have skyrocketed from obscurity to a peak in the autumn of 2025, it is clear this buzzword signifies a tangible and powerful shift in the capabilities of artificial intelligence, moving us toward a world where AI doesn't just advise but actively accomplishes tasks on our behalf.
