<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Sai Chaitanya Bandaru on Medium]]></title>
        <description><![CDATA[Stories by Sai Chaitanya Bandaru on Medium]]></description>
        <link>https://medium.com/@chaitanya21b?source=rss-6efeae78bf4e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*GTzDGyDIj86lG2WH</url>
            <title>Stories by Sai Chaitanya Bandaru on Medium</title>
            <link>https://medium.com/@chaitanya21b?source=rss-6efeae78bf4e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 09:54:37 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@chaitanya21b/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Technical Frameworks for building Agentic AI Systems]]></title>
            <link>https://medium.com/quantrium-tech/technical-frameworks-for-building-agentic-ai-systems-8bb5d9bd23d9?source=rss-6efeae78bf4e------2</link>
            <guid isPermaLink="false">https://medium.com/p/8bb5d9bd23d9</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[crew-ai]]></category>
            <category><![CDATA[phidata]]></category>
            <category><![CDATA[langgraph]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <dc:creator><![CDATA[Sai Chaitanya Bandaru]]></dc:creator>
            <pubDate>Tue, 25 Mar 2025 05:15:30 GMT</pubDate>
            <atom:updated>2025-07-02T14:58:12.598Z</atom:updated>
            <content:encoded><![CDATA[<h4>QUANTRIUM GUIDES</h4><h4><em>Building Scalable and Adaptive AI Agents with Modern Frameworks</em></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*E2xMOjDGtp5ViUerd3aR_w.jpeg" /></figure><p>In our previous article titled <strong><em>“</em></strong><a href="https://medium.com/quantrium-tech/what-is-agentic-ai-and-why-all-the-hype-ec4e5fd43813"><strong><em>What Is Agentic AI and Why All the Hype?</em></strong></a><strong><em>”</em></strong>, we explored the foundational concepts of Agentic AI, its core components, and how it stands apart from other AI paradigms like Generative AI and Traditional AI. We discussed how Agentic AI systems excel in autonomous decision-making, proactive behavior, and adaptability.</p><p>If you haven’t read that article yet, we highly recommend giving it a look before diving into this one. While it’s not a prerequisite, it will provide valuable context to better understand the technical frameworks discussed here.</p><p>As AI technology advances, selecting the right type of AI for a given problem is more crucial than ever. Two key paradigms, <strong>Agentic AI</strong> and <strong>Generative AI</strong>, offer distinct capabilities. Agentic AI excels in autonomous decision-making and goal-oriented tasks, while Generative AI is best suited for content creation, pattern recognition, and synthesis of new ideas. But how do you determine which one is best for your specific needs?</p><p>In this article, we’ll explore technical frameworks that empower developers to build robust Agentic AI systems, with a focus on<strong> PhiData, CrewAI, and LangGraph. </strong>These frameworks provide powerful tools for constructing autonomous agents capable of handling complex workflows and dynamic environments.</p><h3>How to Choose the Right AI?</h3><p>Selecting the right AI system depends on the nature of your problem. Here’s a checklist to help determine whether your use case requires an Agentic AI system.</p><p><strong>Checklist</strong></p><ul><li>Do you need autonomous decision-making?</li><li>Do you require actions to be taken without human intervention?</li><li>Does your system need to adapt to changing tasks and environments?</li><li>Do you need a system capable of handling a dynamic workflow?</li><li>Do you want the AI to proactively perceive and respond to inputs from multiple sources?</li><li>Does your system require real-time adaptation based on incoming data?</li><li>Do you need AI to collaborate with other AI agents or systems to achieve complex goals?</li><li>Are you looking for AI to execute multi-step tasks with minimal human supervision?</li></ul><p>Now, let’s apply this checklist to different scenarios and determine which type of AI is best suited for each case.</p><h4>Case 1: Fully Autonomous IT Incident Resolution System</h4><p>A large enterprise faces frequent IT incidents, such as server failures and network issues. They need a system that can detect, diagnose, and resolve these issues autonomously, without human intervention.</p><p><strong>Checklist</strong></p><ul><li>Do you need autonomous decision-making? ✔ <strong>Yes</strong></li><li>Do you require actions to be taken without human intervention? ✔ <strong>Yes</strong></li><li>Does your system need to adapt to changing tasks and environments? ✔ <strong>Yes</strong></li><li>Do you need a system capable of handling a dynamic workflow? ✔ <strong>Yes</strong></li><li>Do you want the AI to proactively perceive and respond to inputs from multiple sources? ✔ <strong>Yes</strong></li><li>Does your system require real-time adaptation based on incoming data? ✔ <strong>Yes</strong></li><li>Do you need AI to collaborate with other AI agents or systems to achieve complex goals? ✔ <strong>Yes</strong></li><li>Are you looking for AI to execute multi-step tasks with minimal human supervision? ✔ <strong>Yes</strong></li></ul><p>This system requires <strong>Agentic AI</strong>, as it must autonomously detect, diagnose, and resolve IT issues, adapt to new problems, and operate dynamically without human intervention.</p><h3>Case 2: AI-Powered Marketing Content Generator</h3><p>A marketing team wants to automate content creation for social media, blogs, and product descriptions using AI. The AI must generate engaging, high-quality content, but human reviewers will decide which content to use.</p><p><strong>Checklist</strong></p><ul><li>Do you need autonomous decision-making? ❌ <strong>No </strong>(Humans make the final decision)</li><li>Do you require actions to be taken without human intervention? ❌ <strong>No </strong>(AI assists but doesn’t act independently)</li><li>Does your system need to adapt to changing tasks and environments? ❌ <strong>No </strong>(The task — content creation — remains the same)</li><li>Do you need a system capable of handling a dynamic workflow? ❌ <strong>No </strong>(Workflow is structured and requires human oversight)</li><li>Do you want the AI to proactively perceive and respond to inputs from multiple sources? ✔ <strong>Yes </strong>(AI may use different data sources for content)</li><li>Does your system require real-time adaptation based on incoming data? ❌ <strong>No </strong>(Static content generation doesn’t need real-time changes)</li><li>Do you need AI to collaborate with other AI agents or systems to achieve complex goals? ❌ <strong>No </strong>(AI works independently)</li><li>Are you looking for AI to execute multi-step tasks with minimal human supervision? ❌ <strong>No </strong>(Humans oversee the final output)</li></ul><p><strong>Generative AI</strong> is the best fit for this case. The AI assists with content creation but doesn’t act autonomously or adapt dynamically — it supports human creativity rather than replacing decision-making.</p><h3>Case 3: AI-Driven Customer Support Assistant</h3><p>A company wants an AI-powered customer support assistant that can handle FAQs with static responses while also escalating urgent issues autonomously, ensuring a conversational and natural flow.</p><p><strong>Checklist</strong></p><ul><li>Do you need autonomous decision-making? ✔ <strong>Yes</strong> (For issue escalation and handling FAQs)</li><li>Do you require actions to be taken without human intervention? ✔ <strong>Yes </strong>(AI responds automatically and escalates when needed)</li><li>Does your system need to adapt to changing tasks and environments? ❌ <strong>No</strong> (Most responses are predefined, except escalations)</li><li>Do you need a system capable of handling a dynamic workflow? ✔ <strong>Yes</strong> (Customer queries vary, requiring adaptable conversations)</li><li>Do you want the AI to proactively perceive and respond to inputs from multiple sources? ✔ <strong>Yes</strong> (AI needs to process customer queries from multiple channels)</li><li>Does your system require real-time adaptation based on incoming data? ✔ <strong>Yes</strong> (AI must adjust responses based on customer input)</li><li>Do you need AI to collaborate with other AI agents or systems to achieve complex goals? ❌ No (Primarily a single-agent task)</li><li>Are you looking for AI to execute multi-step tasks with minimal human supervision? ✔ <strong>Yes</strong> (Handles conversations and escalations without constant oversight)</li></ul><p>This use case is <strong>partially agentic</strong>. The AI autonomously manages conversations and escalations but still relies on predefined rules for many responses. A <strong>hybrid approach</strong> combining <strong>Agentic AI + Generative AI</strong> is ideal, where Generative AI handles conversational interactions, while Agentic AI manages decision-making for escalations.</p><h3>Final Takeaways</h3><ul><li><strong>Agentic AI</strong> is best for systems requiring <strong>autonomous decision-making, real-time adaptation, and dynamic workflows</strong>.</li><li><strong>Generative AI</strong> excels at <strong>content creation and structured tasks where human oversight is required</strong>.</li><li><strong>Hybrid AI</strong> works well when <strong>both adaptability and creativity are needed</strong>, such as in AI-driven customer support.</li></ul><p>By applying this checklist to different cases, you can determine the right AI approach for your needs.</p><h3>Technical Frameworks for Agentic AI</h3><p>Agentic AI frameworks are gaining traction due to their ability to automate complex tasks with adaptability and decision-making capabilities. This guide covers three prominent frameworks: <strong>PhiData</strong>, <strong>CrewAI</strong>, and <strong>LangGraph</strong>, providing practical insights for implementing AI agents using these tools.</p><p><strong>Phidata (now Agno)</strong>: Phidata, recently rebranded as Agno, is an open-source framework for building multimodal AI agents and workflows. It allows developers to create agents equipped with memory, knowledge, tools, and reasoning capabilities, supporting various data types such as text, images, audio, and video.</p><p><strong>CrewAI</strong>: CrewAI is a lean, Python-based framework designed to orchestrate role-playing AI agents that collaborate to tackle complex tasks. Independent of other agent frameworks, it offers both high-level simplicity and precise low-level control, enabling the creation of autonomous AI agents tailored to diverse scenarios.</p><p><strong>LangGraph</strong>: Developed by LangChain, LangGraph is a low-level orchestration framework designed for building controllable AI agents. It enables the creation of stateful, multi-actor applications with large language models (LLMs), facilitating complex task automation and custom LLM-backed experiences.</p><h3>Hands-on Implementation Guide</h3><p>Regardless of the <strong>Agentic AI framework</strong> you choose, the core implementation steps remain the same. Follow this structured approach to build and deploy your agentic system effectively.</p><p><strong>Step 1: Define the Problem Statement</strong></p><p>Clearly outline the objectives of your AI agent. (Eg: Coordinating multiple agents for data scraping and analysis)</p><p><strong>Step 2: Create an Agent</strong></p><p>Develop the AI agent to execute tasks based on the defined objectives.</p><p><strong>Step 3: Integrate the Necessary Tools</strong></p><p>Equip your agent with the appropriate tools to enhance its functionality and task execution.</p><p><strong>Step 4: Add an LLM (if required)</strong></p><p>If natural language understanding or generation is needed, integrate a <strong>Large Language Model (LLM)</strong> to improve the agent’s capabilities.</p><p><strong>Step 5: Monitor the Agentic System</strong></p><p>Evaluate the initial performance of the system to ensure it functions as expected.</p><p><strong>Step 6: Optimize with Feedback or Human-in-the-Loop</strong></p><p>If improvements are needed, incorporate <strong>feedback loops</strong> or <strong>human oversight</strong> to enhance accuracy and efficiency.</p><h3>Creating a simple finance agent using Phidata</h3><p>Lets create a simple Finance agent supported by YFinance Tool in phidata. This is an <strong>episodic</strong> <strong>task</strong> (that have a terminal state). The results can directly printed on terminal.</p><p><strong>Install Phidata</strong></p><pre>pip install -U phidata==2.7.9</pre><p><strong>Requirements</strong></p><ol><li><strong>LLM </strong>— Add llm keys in env and configure the keys (the below code uses <strong>OpenAI</strong>)</li><li><strong>Tools </strong>— YFinanceTools for Finance agent</li></ol><p><strong>Install Tools</strong></p><p>Install yfinance library for YFinanceTools</p><pre>pip install yfinance==0.2.52</pre><p><strong>Create a Finance Agent</strong></p><p>Create a file finance_agent.py</p><pre>from phi.agent import Agent<br>from phi.model.openai import OpenAIChat<br>from phi.tools.yfinance import YFinanceTools<br><br>finance_agent = Agent( <br>    name=&quot;Finance Agent&quot;, # name for the agent <br>    model=OpenAIChat(id=&quot;gpt-4o&quot;), # llm model<br>    tools=[YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True, company_news=True)],  # tools for the agent<br>    instructions=[&quot;Use tables to display data&quot;],  # list of instructions<br>    show_tool_calls=True, # logs the AI tool calls<br>    markdown=True, # formats in Markdown for structured text<br>)<br>finance_agent.print_response(&quot;Summarize analyst recommendations for NVDA&quot;, stream=True)</pre><p><strong>Run the agent</strong></p><pre>python finance_agent.py</pre><p>This is an episodic task which prints the info on the terminal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tn8t1QgQEZzQVB-hI56ZEg.png" /></figure><h4>Advantages</h4><ul><li><strong>Ease of Building</strong> — Designed for quick and simple development.</li><li><strong>Wide Range of Tools</strong> — Offers a variety of built-in <a href="https://docs.phidata.com/tools/introduction">tools </a>and supports custom tool creation.</li><li><strong>Playground</strong> — Provides an interactive playground for testing and building agents.</li><li><strong>Structured Agent Creation</strong> — PhiData agents have well-defined fields, making them easy to set up.</li><li><strong>Memory Support</strong> — Includes a <a href="https://docs.phidata.com/vectordb/introduction"><strong>vector database</strong></a> for storing and retrieving information.</li><li><strong>Embeddings &amp; Chunking</strong> — Supports efficient data processing and retrieval.</li><li><strong>Multi-Agent Collaboration</strong> — Allows multiple agents to work together.</li></ul><h4><strong>Drawbacks</strong></h4><ul><li><strong>Performance Limitations</strong> — The <strong>PhiData Playground crashes</strong> during heavy, continuous tasks due to load and complexity.</li><li><strong>Limited Visualization</strong> — No built-in visualization tools for better workflow monitoring.</li><li><strong>Working Transparency</strong> — Debugging tools are available, but <strong>understanding dynamic workflows can be difficult</strong> due to limited interpretability.</li></ul><h3>2. CrewAI — Building a Finance Agent</h3><p>Lets create a Agentic AI system with defined tasks for analyzing and reporting the stock market. Since its not a simple task we are going for multi-agent system for better performance and task delegation.</p><p><strong>Install CrewAI</strong></p><pre>pip install crewai==0.108.0</pre><p><strong>Setup API Keys</strong></p><p>CrewAI requires API keys for integrating with:</p><ul><li><strong>LLMs</strong> (e.g., OpenAI, Anthropic, etc.)</li><li><strong>Databases</strong> (for data storage and retrieval)</li></ul><p><strong>Components of a CrewAI Agent</strong></p><ul><li><strong>Role</strong>: Defines the agent’s primary function (e.g., <em>Financial Analyst</em>, <em>Reporting Specialist</em>).</li><li><strong>Goal</strong>: Outlines the agent’s intended outcome or objective (e.g., <em>Analyze NVDA stock performance</em>).</li><li><strong>Backstory</strong>: Provides context about the agent’s expertise, experience, or specialization to guide its behavior.</li><li><strong>Tools</strong>: Specifies the resources, APIs, or services the agent will use to achieve its goal (e.g., <em>YFinance</em>, <em>Google Docs</em>).</li></ul><p><strong>Build a Finance Agent</strong></p><pre>from crewai import Crew, Agent, Task<br>from openai import OpenAI<br><br>finance_agent = Agent(<br>    role=&quot;Financial Analyst&quot;,<br>    goal=&quot;Analyze NVDA stock performance and provide insights&quot;,<br>    backstory=&quot;An experienced financial analyst with expertise in stock market trends, known for delivering precise insights to guide investment strategies.&quot;,<br>    tools=[&quot;YFinance&quot;, &quot;AlphaVantage&quot;],<br>    model=OpenAI(&quot;gpt-4o&quot;),<br>)<br><br>reporting_agent = Agent(<br>    role=&quot;Reporting Specialist&quot;,<br>    goal=&quot;Generate a detailed report on NVDA performance&quot;,<br>    backstory=&quot;A skilled technical writer with extensive experience in financial reporting, adept at transforming data into clear, actionable insights.&quot;,<br>    tools=[&quot;Google Docs&quot;, &quot;Markdown&quot;],<br>    model=OpenAI(&quot;gpt-4o&quot;),<br>)<br><br>analysis_task = Task(<br>    description=&quot;Summarize analyst recommendations for NVDA&quot;,<br>    agent=finance_agent<br>)<br><br>report_task = Task(<br>    description=&quot;Create a structured report with insights&quot;,<br>    agent=reporting_agent<br>)<br><br>crew = Crew(agents=[finance_agent, reporting_agent], tasks=[analysis_task, report_task])<br>crew.kickoff()</pre><p><strong>Run the CrewAI Agent</strong></p><pre>python crewai_agent.py</pre><h4><strong>Observations</strong></h4><ul><li>CrewAI adopts a <strong>collaborative agent model</strong> where multiple agents can coordinate tasks.</li><li>Ideal for <strong>multi-step workflows</strong> that require agents to share information and refine decisions.</li></ul><h4><strong>Advantages</strong></h4><ul><li><strong>Robust Collaboration:</strong> CrewAI excels in <strong>multi-agent orchestration</strong>, enabling teamwork across specialized agents.</li><li><strong>Flexible Workflows:</strong> Supports complex, sequential decision-making processes.</li><li><strong>Human-in-the-Loop Support:</strong> Enables seamless human oversight during crucial decision points.</li><li><strong>Dynamic Role Assignment:</strong> Agents can adapt their roles and switch tasks as objectives evolve.</li></ul><h4>Drawbacks</h4><ul><li><strong>Setup Complexity:</strong> Requires API keys, model configurations, and additional dependencies.</li><li><strong>Learning Curve:</strong> Building multi-agent workflows may require more effort compared to PhiData.</li><li><strong>Execution Overhead:</strong> Managing multiple agents may introduce performance overhead for smaller tasks.</li></ul><h3>3. Creating a chatbot using LangGraph</h3><p>Lets create a chatbot using LangGraph with Anthropic</p><p><strong>Install LangGraph</strong></p><pre>pip install langgraph==0.3.7</pre><p><strong>Requirements</strong></p><p>LLM — requires API keys (used Anthropic for the code below)</p><p><strong>Initialize an LLM</strong></p><p>Create a file langgraph_workflow.py and add the following</p><pre>from langchain_anthropic import ChatAnthropic<br><br>llm = ChatAnthropic(model=&quot;claude-3-5-sonnet-20240620&quot;)</pre><p><strong>Create a LangGraph Workflow</strong></p><p>LangGraph is designed for workflow automation by defining nodes (steps) and edges (transitions) in the process.</p><pre>from typing import Annotated<br>from typing_extensions import TypedDict<br>from langgraph.graph import StateGraph, START, END<br>from langgraph.graph.message import add_messages<br><br>class State(TypedDict):<br>    # Messages have the type &quot;list&quot;. The `add_messages` function<br>    # in the annotation defines how this state key should be updated<br>    # (in this case, it appends messages to the list, rather than overwriting them)<br>    messages: Annotated[list, add_messages]<br><br>def chatbot(state: State):<br>    return {&quot;messages&quot;: [llm.invoke(state[&quot;messages&quot;])]}<br><br>graph_builder = StateGraph(State)<br><br># first arg is the unique node name, second arg is the function or object that <br># will be called whenever the node is used.<br>graph_builder.add_node(&quot;chatbot&quot;, chatbot)<br>graph_builder.add_edge(START, &quot;chatbot&quot;)<br>graph_builder.add_edge(&quot;chatbot&quot;, END)<br><br>graph = graph_builder.compile()</pre><p><strong>Visualize the workflow</strong></p><pre>with open(&quot;data/chat_bot_workflow.png&quot;, &quot;wb&quot;) as f:<br>    f.write(graph.get_graph().draw_mermaid_png())</pre><figure><img alt="" src="https://cdn-images-1.medium.com/max/107/1*4nuHRISDExiwRq_jl6rCMA.png" /><figcaption>Chat bot workflow</figcaption></figure><p><strong>Stream the updates</strong></p><pre>def stream_graph_updates(user_input: str):<br>    for event in graph.stream({&quot;messages&quot;: [{&quot;role&quot;: &quot;user&quot;, &quot;content&quot;: user_input}]}):<br>        for value in event.values():<br>            print(&quot;Assistant:&quot;, value[&quot;messages&quot;][-1].content)<br><br>while True:<br>    try:<br>        user_input = input(&quot;User: &quot;)<br>        if user_input.lower() in [&quot;quit&quot;, &quot;exit&quot;, &quot;q&quot;]:<br>            print(&quot;Goodbye!&quot;)<br>            break<br><br>        stream_graph_updates(user_input)<br>    except Exception:<br>        # fallback if input() is not available<br>        user_input = &quot;What do you know about LangGraph?&quot;<br>        print(&quot;User: &quot; + user_input)<br>        stream_graph_updates(user_input)<br>        break</pre><p><strong>Run the LangGraph Workflow</strong></p><pre>python langgraph_workflow.py</pre><h4>Observations</h4><ul><li>LangGraph is designed for <strong>structured workflow automation</strong>.</li><li>Each node represents a <strong>task</strong> or <strong>decision point</strong>, and edge transitions to control the workflow and also making it easy to visualize complex processes.</li></ul><h4>Advantages</h4><ul><li><strong>Clear Workflow Design:</strong> Nodes and edges simplify creating structured automation pipelines.</li><li><strong>Flexible Execution:</strong> Can dynamically adapt to different data inputs or logic paths.</li><li><strong>Modular Design:</strong> Encourages reusable code through modular task definitions.</li></ul><h4>Drawbacks</h4><ul><li><strong>Limited Tooling Integration:</strong> Compared to PhiData or CrewAI, LangGraph may require more custom code for specific data sources or APIs.</li></ul><h3>Comparative Summary</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/723/1*iXe3DMZxROR6_F3gAEv5dA.png" /></figure><h3>Conclusion</h3><p>This technical deep dive is not aimed at identifying the “best” framework but rather exploring the strengths, limitations, and ideal use cases for each. Each framework excels in different scenarios:</p><ul><li><strong>PhiData</strong> is best suited for simple tasks requiring <strong>quick implementation</strong> and minimal overhead. Its lightweight nature makes it ideal for straightforward, episodic workflows.</li><li><strong>CrewAI</strong> is the optimal choice for <strong>multi-agent collaboration</strong>, especially when tasks demand dynamic coordination, role management, and adaptable workflows.</li><li><strong>LangGraph</strong> stands out for its <strong>visual workflow design</strong> and clear process control, making it ideal for structured automation with defined steps and conditions.</li></ul><p>In essence, there’s no universal “best” framework — the right choice hinges on your project’s complexity, data flow, and adaptability needs. Combining these frameworks strategically can unlock powerful Agentic AI solutions for enterprise-scale applications.</p><p>These observations are based on our research. However, as AI frameworks continue to evolve, future upgrades may impact their capabilities. This technical article may not remain valid over time, and it is recommended to stay updated with the latest advancements.</p><p>This article is co‑authored by Yash Goel and me, combining our expertise in Agentic AI systems. Yash Goel contributed significantly by analyzing technical frameworks and providing insightful comparisons among them.</p><p><strong>References:</strong></p><ol><li><a href="https://www.langchain.com/langgraph?utm_source=chatgpt.com">https://www.langchain.com/langgraph</a></li><li><a href="https://docs.phidata.com/introduction?utm_source=chatgpt.com">https://docs.phidata.com/introduction</a></li><li><a href="https://docs.crewai.com/introduction?utm_source=chatgpt.com">https://docs.crewai.com</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=8bb5d9bd23d9" width="1" height="1" alt=""><hr><p><a href="https://medium.com/quantrium-tech/technical-frameworks-for-building-agentic-ai-systems-8bb5d9bd23d9">Technical Frameworks for building Agentic AI Systems</a> was originally published in <a href="https://medium.com/quantrium-tech">Quantrium.ai</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What Is Agentic AI and Why All the Hype?]]></title>
            <link>https://medium.com/quantrium-tech/what-is-agentic-ai-and-why-all-the-hype-ec4e5fd43813?source=rss-6efeae78bf4e------2</link>
            <guid isPermaLink="false">https://medium.com/p/ec4e5fd43813</guid>
            <category><![CDATA[agentic-workflow]]></category>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[agentic-ai-systems]]></category>
            <category><![CDATA[agentic-ai]]></category>
            <category><![CDATA[generative-ai-use-cases]]></category>
            <dc:creator><![CDATA[Sai Chaitanya Bandaru]]></dc:creator>
            <pubDate>Wed, 19 Mar 2025 03:01:48 GMT</pubDate>
            <atom:updated>2025-07-02T14:52:51.538Z</atom:updated>
            <content:encoded><![CDATA[<h4>QUANTRIUM GUIDES</h4><h4>Understanding the Rise of Autonomous AI Systems</h4><figure><img alt="" src="https://cdn-images-1.medium.com/proxy/1*a0GI5MVJwkiW_Q6Sbwp6Bw.jpeg" /><figcaption>Agentic AI</figcaption></figure><p>The recent surge in interest around Agentic AI stems from its potential to revolutionise various industries by automating complex processes and enhancing efficiency. For instance, companies like ServiceNow and Salesforce are developing AI agents to autonomously handle tasks in customer service, HR, IT, and marketing, thereby increasing productivity and reducing operational costs.</p><p>Similarly, in healthcare, AI agents are being utilised to manage administrative tasks, assist in clinical trials, and support physicians in patient care, addressing challenges such as physician burnout and staffing shortages.</p><p>Since the start of 2025, there’s been a surge in hype around <strong>Agentic AI</strong>, along with plenty of speculation about whether AI could eventually replace humans. But what exactly is Agentic AI, and how does it differ from Generative AI? Let’s explore.</p><h3>What’s Agentic AI?</h3><p>Agentic AI refers to autonomous systems capable of dynamic learning, interacting within complex environments, and making decisions independently. Unlike Generative AI, which primarily creates content based on input, Agentic AI can act proactively, take initiative, and adapt to evolving situations without constant human guidance.</p><h3>Agentic AI vs. Other AI Paradigms</h3><p>Here we have compared the three generations of AI depicting evolution of AI over time:</p><ol><li>Agentic AI (3rd Gen),</li><li>Generative AI (2nd Gen)</li><li>Traditional AI (1st Gen)</li></ol><h4>Agentic AI vs Generative AI vs Traditional AI</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*etzDkkIcNJu3FGFYsyA9xg.png" /><figcaption>Agentic AI vs Generative AI vs Traditional AI</figcaption></figure><p>In simple terms:</p><ul><li><strong>Agentic AI</strong> → Proactive, autonomous, decision-making.</li><li><strong>Generative AI</strong> → Content creation, responsive, dependent on prompts.</li><li><strong>Traditional AI</strong> → To solve tasks based on specific rules and pre-defined logic.</li></ul><p>This ability to independently navigate and respond to complex environments is what sets Agentic AI apart and why it’s generating so much excitement.</p><h3>How is Agentic AI different from Generative AI?</h3><p>While Agentic AI promises autonomous decision-making, <strong>LLM + API setups</strong> (alongside RAG and similar architectures) already deliver immense value as highly capable <strong>Generative AI systems</strong>. Here is a detailed comparison.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/725/1*24YBjsna4fQLaqBQcS5mDA.png" /><figcaption>Agentic AI System vs LLM+API System</figcaption></figure><h4>Example: Research Assistant Task</h4><p><strong>Agentic AI:</strong></p><ul><li>One agent autonomously finds relevant sources, another summarizes the content, and a third fact-checks for accuracy.</li><li>The agents work together, refining outputs and continuously improving the final result — all without constant human intervention.</li></ul><p><strong>LLM + API:</strong></p><ul><li>The user manually requests information, summarizes the findings, and verifies accuracy step by step.</li><li>Each stage requires explicit API calls and human oversight, making the process more hands-on and time-consuming.</li></ul><h4>When to Use Each Approach</h4><p><strong>Use Agentic AI when:</strong></p><ul><li>You need <strong>autonomous, multi-step decision-making.</strong></li><li>Tasks require <strong>collaboration among AI agents.</strong></li><li>Adaptability and <strong>self-improvement</strong> are important.</li></ul><p><strong>Use LLM + API when:</strong></p><ul><li>You need <strong>simple, transactional AI responses.</strong></li><li>The task is <strong>well-defined and does not require adaptability.</strong></li><li>You can manually <strong>orchestrate API calls</strong> efficiently.</li></ul><h3>Components of Agentic AI</h3><p>Agentic AI systems have several interconnected components that work together to enable intelligent behavior. Each component plays a crucial role in the overall functioning of the AI system, and they must interact seamlessly to achieve desired outcomes.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/727/1*ldOSQLjYA3vZmgLXt1Tn6w.png" /><figcaption>Components of Agentic AI</figcaption></figure><p><strong>Perception Module</strong> — Captures and interprets real-world data using AI techniques like vision and NLP.</p><p><strong>Cognition Module</strong> — Analyzes information, makes decisions, and plans actions based on learned patterns.</p><p><strong>Action Module</strong> — Executes decisions by performing digital or physical tasks autonomously</p><h3>Why Is There So Much Hype Around Agentic AI?</h3><p>The buzz around Agentic AI isn’t just noise — it’s driven by the tangible benefits these systems bring to organizations. Let’s break down the key factors fueling its rising popularity:</p><ul><li><strong>Automation Needs:</strong> Companies are eager to streamline operations and minimize manual effort, making autonomous systems a powerful asset for boosting efficiency.</li><li><strong>Adaptability:</strong> Agentic AI can adjust to evolving environments and shifting data, ensuring it stays relevant and effective in a variety of scenarios.</li><li><strong>Enhanced Decision-Making:</strong> These systems can analyze complex data and make informed decisions without human intervention, helping businesses act faster and more strategically.</li><li><strong>Scalability:</strong> Agentic AI handles growing workloads with ease, scaling operations without a linear increase in costs or resource demands.</li><li><strong>Cost-Effectiveness:</strong> By automating repetitive tasks and reducing reliance on human oversight, organizations can cut operational expenses while maintaining high performance.</li></ul><p>With these capabilities, it’s easy to see why businesses across industries are eager to explore Agentic AI. But what about its impact on our day-to-day lives?</p><h3>How Can Agentic AI Make Our Lives Easier?</h3><p>Agentic AI isn’t just a futuristic concept — it’s a game-changer for simplifying tasks, boosting productivity, and freeing us up to focus on what truly matters. Here’s how it can transform both work and personal life:</p><ul><li><strong>Automate Workflows:</strong> Seamlessly handle repetitive tasks, from scheduling meetings to managing customer inquiries, freeing up valuable time for more meaningful work.</li><li><strong>Focus on High-Impact Work:</strong> By handling time-consuming, low-value tasks, Agentic AI allows people to concentrate on strategic, creative, and high-priority initiatives.</li><li><strong>Generate &amp; Refine Content:</strong> Create, edit, and optimize content effortlessly — whether it’s drafting reports, crafting emails, or refining presentations.</li><li><strong>Intelligent Data Gathering &amp; Analysis:</strong> Collect and process data from various sources (like voice commands, text, or even receipts) to make faster, more accurate decisions — sometimes even outperforming humans in complex scenarios.</li></ul><p>From personal assistants to workplace optimization, Agentic AI has the potential to revolutionize how we manage time and resources. However, as with any powerful technology, it comes with its own set of challenges.</p><h3>Can We Trust Agentic AI Completely?</h3><p>While Agentic AI holds incredible promise, it’s not without its limitations. Understanding these potential pitfalls is essential for organizations to adopt the technology responsibly.</p><ul><li><strong>The Illusion of Control:</strong> Enterprises may feel in control when setting up AI agents, but many of these systems operate as black boxes. Without proper oversight, AI-driven decisions can become unpredictable — and in highly regulated sectors like finance or healthcare, even small missteps can have serious consequences.<br>😇 <em>Implement robust monitoring and feedback loops to ensure continuous human oversight and mitigate unexpected outcomes.</em></li><li><strong>The Maintenance Challenge:</strong> Although Agentic AI systems promise rapid deployment and adaptive learning, they often lack robust maintenance frameworks. Most automation work happens after deployment, and small process tweaks can cause unintended side effects, turning continuous improvement into a complex challenge.<br>😇<em> Establish clear update and retraining protocols to ensure the AI evolves safely and aligns with changing business needs.</em></li><li><strong>The Reliability Mirage:</strong> Even with error rates as low as 1–5%, inaccuracies in high-stakes environments like banking or legal compliance are unacceptable. A single misstep can lead to financial losses, reputational damage, or legal complications, making complete trust in AI a risky bet.<br>😇 <em>Use AI as an assistive tool for decision support rather than a fully autonomous solution in high-risk areas.</em></li><li><strong>The Governance Gap:</strong> The rise of “shadow AI” — unauthorized or fragmented AI use across organizations — mirrors past issues with shadow IT. Without centralized governance, businesses face compliance breaches, inconsistent processes, and a fragmented tech landscape that’s difficult to manage.<br>😇 <em>Create a centralized AI governance team to set policies, track AI usage, and ensure compliance across the enterprise.</em></li><li><strong>The Learning Dilemma:</strong> Agentic AI lacks structured lifecycle management, which can cause processes to become outdated over time. Without a clear change management strategy, organizations risk AI systems making decisions based on obsolete information, diminishing long-term effectiveness.<br>😇 <em>Implement lifecycle management practices, including regular audits and periodic retraining, to keep AI systems aligned with evolving goals and data.</em></li></ul><h3>Final Thoughts</h3><p>Agentic AI is undeniably powerful technology that promises to reshape industries, streamline workflows, and even make our daily lives more convenient. But with great power comes great responsibility. For businesses and individuals alike, balancing excitement with caution is key.</p><p>By understanding both the potential and the limitations of Agentic AI, we can harness its strengths while mitigating its risks. With the right mix of innovation, governance, and continuous oversight, Agentic AI can be an extraordinary tool for progress.</p><p>So, can we trust Agentic AI completely? Maybe not yet. But can we leverage it to build a smarter, more efficient future? Absolutely.</p><p>This article is co‑authored by Yash Goel and me, bringing together our shared expertise in Agentic AI research. In particular, Yash Goel played a key role in identifying real-world use cases, and providing thoughtful comparisons between different AI paradigms which were essential to shaping the article’s core insights.</p><p><strong>References</strong></p><ol><li><a href="https://www.youtube.com/watch?v=xOS0BhhdUbo">https://www.youtube.com/watch?v=xOS0BhhdUbo</a></li><li><a href="https://www.uipath.com/ai/agentic-ai">https://www.uipath.com/ai/agentic-ai</a></li><li><a href="https://aisera.com/blog/agentic-ai/">https://aisera.com/blog/agentic-ai/</a></li><li><a href="https://www.automationanywhere.com/rpa/agentic-ai#:~:text=Despite%20their%20autonomy%2C%20agentic%20AI,judgment%20or%20deep%20contextual%20understanding">https://www.automationanywhere.com/rpa/agentic-ai#:~:text=Despite%20their%20autonomy%2C%20agentic%20AI,judgment%20or%20deep%20contextual%20understanding</a>.</li><li><a href="https://scet.berkeley.edu/the-next-next-big-thing-agentic-ais-opportunities-and-risks/">https://scet.berkeley.edu/the-next-next-big-thing-agentic-ais-opportunities-and-risks/</a></li><li><a href="https://markovate.com/blog/agentic-ai-architecture/">https://markovate.com/blog/agentic-ai-architecture/</a></li><li><a href="https://nexustrade.io/blog/you-are-an-absolute-moron-for-believing-in-the-hype-of-ai-agents-20250111">https://nexustrade.io/blog/you-are-an-absolute-moron-for-believing-in-the-hype-of-ai-agents-20250111</a></li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec4e5fd43813" width="1" height="1" alt=""><hr><p><a href="https://medium.com/quantrium-tech/what-is-agentic-ai-and-why-all-the-hype-ec4e5fd43813">What Is Agentic AI and Why All the Hype?</a> was originally published in <a href="https://medium.com/quantrium-tech">Quantrium.ai</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>