The SWE Pune Affiliate marked a significant return with the Prajna webinar series, “Establish | Execute | Expand.”
Representing the SWE Pune Affiliate, Prerna Amarnani hosted the first Prajna webinar series session. She graciously set the stage with an overview of SWE’s mission to champion women engineers and foster a culture of growth and inclusion.
The event began with an engaging session by Shweta Mani Ashok, AI cloud development security operations (DevSecOps) architect at Capgemini, who spoke on “The Dawn of Agentic AI.” Her talk explored the architectural foundations of AI systems, offering deep insights into how intelligent agents are designed and deployed in cloud environments.
The session opened with a powerful exploration of what defines truly intelligent systems, including the history and flow of AI from machine learning, to deep learning, to generative AI, to now agentic AI.
Shweta laid the foundation by outlining the core characteristics of agentic AI: autonomy, goal-oriented behaviors, complex decision-making, high adaptability, and environmental perception. These traits mark a shift from traditional rule-based automation to systems capable of independent action, strategic thinking, and dynamic interaction with their surroundings.
As Shweta explained, agentic AI represents a new frontier, where machines not only execute tasks but also understand context, pursue objectives, and evolve with experience.
The next insight offered was a clear distinction between three evolving AI paradigms: generative AI, agent AI, and agentic AI. While often conflated, each serves a unique purpose.
Generative AI focuses on creation — producing content, code, or designs from data. Agent AI is designed to act, executing tasks within defined parameters.
Agentic AI, however, is built to achieve — operating autonomously, pursuing goals, making complex decisions, and adapting to dynamic environments. Shweta emphasized that agentic AI isn’t just a more advanced version of the others — it’s a transformative leap toward machines that can think, plan, and succeed independently.
In the third segment of her talk, Shweta spotlighted the real-world impact of agentic AI across a diverse range of industries. From health care, where autonomous systems can assist in diagnostics and personalized treatment plans, to manufacturing, where intelligent agents optimize production lines and predict equipment failures, the potential is transformative.
In finance, agentic AI enables dynamic risk assessment and fraud detection, while in logistics, it enhances supply chain efficiency through adaptive routing and inventory management. Education benefits from personalized learning experiences powered by goal-driven AI tutors, and smart cities leverage these systems for traffic management, energy optimization, calamities management, and urban planning. Shweta’s examples illustrated how agentic AI is already reshaping the way industries operate, with autonomy and adaptability at the core.
Moving forward, Shweta delved into the technical backbone of agentic AI, outlining the layered architecture that enables these systems to function with autonomy and intelligence. She introduced several agentic AI frameworks, such as LangChain, AutoGPT, and LangGraph, that developers can leverage to build agentic systems today. These tools offer modular capabilities for chaining tasks, managing memory, and integrating external APIs, making it easier than ever to prototype and deploy intelligent agents across domains.
Shweta outlined key challenges in building agentic AI, including defining clear objectives, handling real-world uncertainties, debugging autonomous behavior, and ensuring ethical decision-making. She emphasized that these hurdles can lead to unpredictable or unsafe outcomes if not addressed thoughtfully. To overcome them, she recommended using goal-alignment frameworks, simulation environments, and explainable AI techniques.
In conclusion, she offered a forward-looking perspective on the future of agentic AI, emphasizing its potential to reshape industries, redefine human-machine collaboration, and unlock new levels of autonomy and intelligence.
SWE Pune Affiliate extends its sincere appreciation to Shweta for her compelling session. Her expertise and forward-looking perspective provided valuable clarity on a rapidly evolving field. We thank her for her contribution and look forward to continued engagement.
Make meaningful connections with STEM peers, mentors, allies, and employers at WE Local Pune, SWE’s inspiring local conference that takes place December 4-5, 2025, in Pune, India! Get all the details and reserve your spot on the WE Local Pune website.
