Topics In Demand
Notification
New

No notification found.

Generative versus Agentic AI: A Tale of Two Approaches
Generative versus Agentic AI: A Tale of Two Approaches

32

0

‘It was the age of creation, it was the age of execution; it was the canvas of imagination, it was the machinery of action; it was the dawn of limitless possibilities, it was the twilight of human control — Generative and Agentic AI stood divided yet entwined in their purpose.’

In the over 5,000 years of recorded human history, nothing has come even remotely close to the interest generated around generative Artificial Intelligence (GenAI). Over the last couple of years, a range of tools, led by ChatGPT, have captured the spotlight, revolutionizing content generation, data summarization, and analytical workflows.

From crafting marketing campaigns to distilling complex financial or legal documents, GenAI has proven its value across diverse business use cases, especially in streamlining back-office functions. However, in the high-stakes, operational world of Industrial usage like Oil & Gas (O&G), where uptime, safety, and efficiency are non-negotiable, GenAI’s impact has continued to be minimal for now.

For O&G CXOs, there is a need for AI to be able to tackle the heavy lifting of operations, not just assist with reports generation and recommendations. A popular quip sums up the feeling (almost) – “I want AI to do my household chores while I do art, creative writing and not the other way around.”

Agentic AI: The Transformative AI Enterprises Have Been Waiting For?

Enter Agentic AI, a paradigm shift that promises to bridge the gaps of GenAI. Unlike GenAI, which excels at generating outputs for human review, Agentic AI autonomously makes decisions, executes tasks, and integrates with physical systems in real time. This capability offers the potential to unlock unprecedented efficiency, safety, and cost savings for the global O&G sector.

To understand the potential of Agentic AI in greater depth, consider the fundamental differences between the two AI paradigms. GenAI leverages large language models (LLMs) to create content, think detailed reports, synthetic data, or even conceptual designs, based on user prompts. Agentic AI, on the other hand, builds upon LLMs as a reasoning and decision-making engine, augmented with real-time data integration, planning algorithms, and system interactions across IoT sensors, robotics, or APIs. While GenAI generates ideas, Agentic AI acts on them, driving operational outcomes with minimal human oversight.

A Daily Life Example: Planning a Healthy Dinner

Imagine asking, “Plan a healthy dinner for me tonight.” A GenAI tool would respond with a creative suggestion: “How about a grilled chicken salad with avocado, cherry tomatoes, and a lemon-tahini dressing? Here is a recipe...”

It does present you with a detailed idea (and avoids that eternal conundrum of what to eat – hopefully), but the planning and execution is on you.

Agentic AI, by contrast, acts.

It analyzes your dietary preferences, location, and real-time data, then responds: “I recommend a grilled chicken salad with avocado, tailored to your low-carb preference. I have also located fresh ingredients at a Tesco Express 5 minutes from your office in Central London, which is open until midnight. I can order via Deliveroo for delivery to your hotel by 8 PM, and I will synchronize a recipe app to guide your preparation. Shall I proceed?”

This actionable, integrated approach is what sets Agentic AI apart.

Industrial Example: Optimizing Energy Usage in an Oil Refinery

Now, let us scale this to an O&G context with the prompt: “Optimize energy usage in my refinery this week.” A GenAI system would generate a theoretical plan: “Consider reducing HVAC usage during low-production hours and shifting high-energy processes to off-peak times, such as 2 AM to 5 AM, when electricity rates are lower. Here is a sample schedule: HVAC at 50% from 10 PM to 6 AM.”

The Agentic AI tool, however, would take control of the request. It would integrate with your refinery’s SCADA systems, analyze real-time data from sensors, weather forecasts, and local energy pricing, and reply: “I have analyzed your refinery’s energy consumption patterns and local grid data. To optimize energy this week, I will reduce HVAC usage by 15% during low-production hours (10 PM–6 AM) and reschedule crude distillation unit operations to off-peak hours (2 AM–5 AM), saving 12% on energy costs. I have also adjusted the control systems on these parameters, addressing data latency challenges by leveraging edge computing to process sensor data locally, ensuring real-time responsiveness. I will continue to monitor performance daily for further optimization, mitigating model drift by retraining algorithms with fresh data. Expected savings: $150,000 this week, based on a pilot study by a major Gulf Coast refinery operator [Ref: SPE-2024-1234]. Would you like a detailed report?”

This level of autonomy and integration directly impacts your bottom line.

Interested in learning how this affects you? Click here for part 2 of this two-part series.


That the contents of third-party articles/blogs published here on the website, and the interpretation of all information in the article/blogs such as data, maps, numbers, opinions etc. displayed in the article/blogs and views or the opinions expressed within the content are solely of the author's; and do not reflect the opinions and beliefs of NASSCOM or its affiliates in any manner. NASSCOM does not take any liability w.r.t. content in any manner and will not be liable in any manner whatsoever for any kind of liability arising out of any act, error or omission. The contents of third-party article/blogs published, are provided solely as convenience; and the presence of these articles/blogs should not, under any circumstances, be considered as an endorsement of the contents by NASSCOM in any manner; and if you chose to access these articles/blogs , you do so at your own risk.


L&T Technology Services

© Copyright nasscom. All Rights Reserved.