From Automation to LLMs:
My AI Journey

A year ago, I walked into Xoriant as a fresher, excited to apply what I had learned. My head was full of Python, Machine Learning, and Deep Learning concepts. More than anything, I was passionate about automation, the idea of teaching a machine to handle repetitive tasks so people could focus on more creative work. 

Like many in tech, I have watched the recent explosion of Generative AI with fascination. It’s the topic of every tech blog and team meeting. But for me, this isn’t just a trend. It feels like a natural extension of the journey I started right here, on my very first projects. 

Early Days: Seeing Opportunity Everywhere 

When I first joined, I spent a lot of time observing. I saw many processes and workflows that were being handled manually. My fresher mindset immediately kicked in, and I kept thinking, "We could automate that!" or "A simple script could handle this." 

Of course, I knew I didn't have full perspective. As a new joiner, I didn't understand all the complexities or the reasons why things were done in a certain way. But that initial instinct to find and fix inefficiencies stuck with me. It taught me a valuable lesson: before you can build complex AI, you must first appreciate the power of simple, effective automation. 

My First "Win": The Power of Starting Small 

My chance to apply this thinking came when I got to work on automating an approval flow. It wasn't a glamorous, headline-grabbing AI project. It was a practical, necessary task that was consuming valuable time. 

For every request, end user had to: 

1. Stop what they were doing. 
2. Log into a separate application. 
3. Find the specific item. 
4. Action it manually. 

It was a small interruption, but it happened multiple times a day, breaking their focus and slowing down the entire process. 

The Solution:

We decided to bring the decision to the user. We built a system where the approval and rejection options were embedded directly into the notification email. Now, instead of navigating complex menus, a manager could approve or reject a request with a single click, right from their inbox.

The impact was immediate. We reduced manual effort, minimized errors, and dramatically sped up the entire workflow. This experience was my "aha!" moment. It proved that you don’t need a massive, complex project to make a real impact. Solving a small, tangible problem removing that one annoying, time-wasting step is the perfect first step. It builds confidence, demonstrates value, and creates a foundation for more advanced work. 

Stepping into the World of LLMs 

That foundation in automation prepared me for my next challenge, which threw me right into the world of Large Language Models (LLMs). I was assigned to a project focused on Entity Extraction. 

To put it simply, we were teaching an AI to read through large amounts of text and act like a highly intelligent search function, pulling out specific, crucial pieces of information like: 

• Names 
• Dates 
• Invoice numbers 
• Product codes 

This is a task that would take a human being hours of tedious reading, but an LLM can do it in seconds. 

Working on this project was a game changer for me. It took the theoretical concepts of AI and made them real. I saw firsthand how an LLM wasn't just a chatbot that could write a poem; it was a powerful tool that could be trained on specific business data to solve a very real business problem. It connected the dots between my early interest in automation and the massive potential of Generative AI. We weren't just making a process faster; we were unlocking the value hidden inside unstructured data. 

My One-Year Perspective 

Looking back, my journey from a fresher learning Python to working on an LLM project has shown me a clear path for adopting new technology. You don't have to jump straight to the most complex solution. 

My advice to anyone, whether you’re a fresher or a seasoned pro, is this: 

1. Start by observing. Look for the manual, repetitive tasks around you. 
2. Automate what you can. Solving a simple problem builds momentum. 
3. Leverage that experience. Use your automation mindset to tackle bigger challenges with advanced tools like LLMs. 

The hype around Generative AI is real, but it's not magic. It's a technology that, when applied thoughtfully, can solve problems at every level. And you don’t need a decade of experience to start making a difference. Sometimes, a fresh perspective and a passion for automation are all you need. 

This blog is part of ThoughtForce, an initiative by Xoriant to showcase insights from its House of XFactors, driving thought leadership through collective expertise.