22:36
What is Agentic AI? Important For GEN AI In 2025
Krish Naik
Overview
This video introduces agentic AI, differentiating it from generative AI and highlighting its significance for the future of AI development, particularly in 2025. Generative AI focuses on content creation through LLMs and prompt engineering. Agentic AI, however, involves autonomous AI systems designed to achieve specific goals by executing complex workflows independently, often integrating multiple tools and other AI agents. The video discusses key frameworks like Fidd, Microsoft Autogen, Langflow, and LangGraph for building agentic applications and demonstrates a practical example of a financial analyst agent.
How was this?
Save this permanently with flashcards, quizzes, and AI chat
Chapters
- Generative AI, prominent in 2024, focuses on content creation using Large Language Models (LLMs) and prompt engineering.
- Agentic AI, emerging as a key trend for 2025, refers to autonomous AI systems capable of independent task execution.
- The core difference lies in the goal: generative AI creates content, while agentic AI achieves specific business outcomes through complex, independent workflows.
- Agentic AI systems can self-tune and improve performance over time, a capability not present in traditional generative AI.
Understanding the shift from generative to agentic AI is crucial for anticipating future AI capabilities and developing relevant skills for the evolving job market.
Generative AI: Asking an LLM to write a poem about a topic. Agentic AI: An AI system that researches stock market data, analyzes news, and provides a buy/sell recommendation.
- Agentic AI systems operate autonomously, working independently without constant human intervention to achieve a defined goal.
- They can integrate and utilize various external tools (like search engines, APIs, or databases) to gather information and perform actions.
- Frameworks like LangChain enable the integration of these tools, allowing LLMs to access real-time or specialized data beyond their training set.
- Agentic AI can orchestrate complex workflows by chaining multiple tools and even other AI agents together.
This autonomous capability and tool integration allow AI to tackle more complex, real-world problems that require multi-step processes and access to dynamic information.
A user asks an agentic AI to compare Tesla and Nvidia stocks. The AI uses a tool like YT Finance to get stock data, another tool to fetch recent financial news, and then an LLM to analyze and synthesize this information to provide a recommendation.
- Several open-source frameworks facilitate the development of agentic AI applications.
- Fidd allows the creation of specialized agents (e.g., financial, legal) that can be executed in workflows and integrated with any LLM.
- Microsoft Autogen provides a framework for building agentic AI applications, enabling conversational interactions between agents.
- Langflow offers a no-code, drag-and-drop interface for designing complex agentic workflows and generating code.
- LangGraph is highlighted for its ability to create highly complex workflows and integrate multiple agents, models, and data preprocessing techniques.
These frameworks democratize the creation of sophisticated AI agents, enabling developers to build powerful applications without starting from scratch.
Using Langflow's visual interface to drag and drop different agent components, connect them to tools like a search engine, and define the sequence of operations for a research task.
- The video demonstrates an agentic AI application built using Fidd, acting as a financial analyst.
- This agent integrates a web search agent (using DuckDuckGo) for recent news and a financial agent (using YT Finance) for stock data.
- The application allows users to ask complex financial questions, such as stock recommendations between two companies.
- The agent autonomously executes a workflow, gathering data from multiple sources, analyzing it, and providing a synthesized recommendation.
This hands-on example illustrates the practical power of agentic AI in solving real-world business problems by automating complex analytical tasks.
Asking the demo agent, 'Can you suggest me which stock to buy, is it between Tesla and Nvidia?' and receiving a detailed analysis and recommendation based on current data and news.
Key takeaways
- Agentic AI represents a significant advancement beyond generative AI, focusing on autonomous goal achievement through complex workflows.
- The ability of agentic AI to integrate and utilize external tools is key to its power in accessing and processing real-time information.
- Autonomous execution and self-improvement are defining characteristics that set agentic AI apart.
- The development of agentic AI is being accelerated by accessible open-source frameworks like Fidd, Autogen, Langflow, and LangGraph.
- Agentic AI has the potential to disrupt industries by automating complex tasks and decision-making processes.
- Learning to build and implement agentic AI applications will be a valuable skill set for developers in the coming years.
Key terms
Agentic AIGenerative AIAutonomous AI SystemLLM (Large Language Model)Prompt EngineeringWorkflowTools (in AI context)Frameworks (AI development)FiddMicrosoft AutogenLangflowLangGraph
Test your understanding
- What is the fundamental difference between generative AI and agentic AI in terms of their primary function?
- How does agentic AI leverage external tools to overcome the limitations of traditional LLMs?
- Explain the concept of an autonomous AI system within the context of agentic AI.
- Why are frameworks like Fidd and Autogen important for the development and adoption of agentic AI?
- Describe a hypothetical complex workflow that an agentic AI system might execute to achieve a business outcome.