While AI-powered chatbots quickly captured the public’s attention, a more powerful application of generative AI is currently making waves among business leaders. This technology, known as agentic AI, is generating considerable excitement.
“Agentic AI isn’t just another trend; it represents a fundamental shift in AI capabilities,” tech expert Bernard Marr wrote in his Intelligence Revolution newsletter on Monday.
At its core, agentic AI refers to autonomous AI systems that can act independently to achieve specific goals. Unlike traditional AI models that merely respond to commands or carry out predefined tasks, agentic AI is capable of making decisions, planning actions, and learning from experience to pursue objectives set by human creators.
“Agentic AI is currently one of the most talked-about developments,” noted Jason Wong, vice president and analyst at Gartner, a research and advisory firm in Stamford, Connecticut.
He elaborated that this technology doesn’t just understand intent and perform simple tasks like retrieving information; it can also take action. “It could retrieve an API, use a tool, or even generate code, like writing Python code to solve a problem,” Wong explained to TechNewsWorld.
The key feature of agentic AI is its ability to plan and solve problems by activating tools to address specific questions or tasks, he added.
Going Beyond Generative AI
Scott Dylan, founder of NexaTech Ventures, a venture capital firm in Manchester, UK, argued that agentic AI goes much further than generative AI. “While generative AI focuses on creating content — such as text, images, or code — from existing data, agentic AI possesses autonomy,” he told TechNewsWorld. “It can make decisions, act, and adapt in real time without ongoing human input.”
“Imagine evolving from a tool that merely offers suggestions to one that independently carries out tasks and learns from its environment,” he said.
Agentic AI is an important advancement over traditional generative AI by adding self-directed reasoning, dynamic computation, and adaptive problem-solving, according to Dev Nag, CEO and founder of QueryPal, an enterprise chatbot in San Francisco.
“Unlike generative AI, which mainly focuses on content creation based on prompts, agentic AI can independently allocate time to more complex tasks, use hidden thought processes, and apply reinforcement learning to optimize its decision-making,” Nag explained to TechNewsWorld.
“This development enables agentic AI to handle more complex challenges and adjust its methods based on the task, expanding its abilities beyond simple text generation to more human-like problem-solving across diverse data fields,” he said. “It’s clear that modern agentic AI — like OpenAI’s o1 — builds on generative AI as a foundation but can achieve a broader range of objectives.”
Transformative Technology
The powerful capabilities of agentic AI have the potential to revolutionize numerous industries.
“Agentic AI can transform industries by automating not only repetitive tasks but also complex decision-making processes. For instance, in supply chain management, agentic AI could predict and respond to disruptions in real time, optimizing routes and inventory without human intervention,” said Hodan Omaar, senior AI policy analyst at the Center for Data Innovation, a think tank in Washington, D.C. that focuses on data, technology, and public policy.
“Businesses are on the verge of a significant shift due to agentic AI,” added Scott Dylan, founder of NexaTech Ventures. “It’s not just about automating tasks; it’s about empowering systems to make complex decisions. In finance, this level of autonomy could lead to more personalized customer service and fraud prevention systems that evolve with emerging threats without the need for constant human oversight.”
Dylan also highlighted the technology’s potential in healthcare. “Imagine a healthcare system that doesn’t just diagnose based on symptoms, but actively monitors patients post-diagnosis, adapting treatment plans as it learns from continuous data. While this is a long-term vision, agentic AI is laying the foundation for that reality.”
Dev Nag, CEO of QueryPal, pointed out that agentic AI could revolutionize industries such as law, medicine, and finance by automating complex cognitive tasks. While it might replace jobs involving routine analysis, it would also create new roles centered around AI oversight and human-AI collaboration.
“Agentic AI’s ability to scale in real-time to solve increasingly difficult problems without needing larger models or more data could democratize access to advanced AI tools, enabling smaller businesses to take advantage of powerful capabilities,” Nag explained.
“This new paradigm of runtime scaling offers a novel approach to AI development, going beyond just scaling hardware or training data, which has been the focus of AI companies over the past two years,” he added.
Shared Brain, Shared Problems
Like generative AI, agentic AI faces its own set of challenges. “Because AI agents use a language model as their ‘brain,’ they inherit all the issues of generative AI, plus some additional ones,” said Sandi Besen, an applied AI researcher at IBM and Neudesic, a global professional services firm.
“Moreover, when multiple agents are used together and are given the ability to collaborate, the inherent variability in generative AI is amplified,” Besen noted. “However, it is possible to mitigate these issues by ensuring proper evaluation and maintaining human oversight in the AI system.”
“Agentic AI, like other forms of AI, has the potential to significantly boost productivity. By handling the multiple steps involved in many tasks, it can automate work and save users time and money,” said David Inserra, fellow for free expression and technology at the Cato Institute, a think tank based in Washington, D.C.
“While some may inevitably misuse such AI tools or use them to create harmful content, the many positive applications of this technology mean it should be allowed to develop freely, without excessive government regulation like that seen in the EU,” Inserra argued. “Due to such regulations, major tech companies are already withholding new AI tools in Europe, which ultimately puts Europeans at a disadvantage.”
Closer to AGI?
Given that agentic AI allows generative AI to take action, does this bring us any closer to the long-sought goal of artificial general intelligence (AGI) and truly thinking machines?
“A key feature of general intelligence, whether in humans or animals, is the ability to adapt — sensing environmental cues, reacting to them, and learning from these reactions. In this sense, agentic AI represents a small but significant step toward general intelligence,” said Rogers Jeffrey Leo John, co-founder and CTO of DataChat, a no-code generative AI platform for analytics in Madison, Wisconsin.
“However,” he cautioned, “we are still far from achieving true general intelligence, which would involve applying knowledge gained in one situation to entirely different contexts.”
Shawn DuBravac, CEO and president of the Avrio Institute, a technology consulting firm based in Madison, was more skeptical about agentic AI being a step toward AGI. “I would argue that agentic AI is not a precursor to AGI,” he told TechNewsWorld. “It’s not clear that we will reach AGI through a linear progression from current AI technologies like agentic AI.”
“In fact, I think it’s unlikely,” DuBravac added. “If we do reach AGI, I believe it will involve breakthroughs and new paradigms of intelligence that are fundamentally different from what we’ve achieved so far or what we can expect in the near future.”
