Tinkering, Innovation, and Automation: The raggity.ai Story : Part 1

My Journey of Tinkering and Problem-Solving

I’ve always had a passion for tinkering and solving problems. My very first invention dates back to the early 2000s when I built an aux cord for my Sony boombox. I rewired the tape deck—converting an RCA output to a 1/8-inch phono jack—so I could plug in my mini disk player. In hindsight, using a cassette tape converter might have been easier, but for a 10-year-old, that DIY project felt like a bold, visionary experiment. That same spirit of deconstructing, understanding, and rebuilding has fueled every project I’ve tackled since.

The Spark: Discovering AI and the Magic of Vector Databases

Fast forward a few years—I found myself immersed in the world of AI at Deep Atlas’s machine learning bootcamp. I was floored by the concepts of vector space and the step function benefits of vector databases. These systems unlock a level of data usability and time savings that traditional workflows simply can’t match.

To put it simply, with vector search there’s no need to tag or structure data manually. Everything gets transformed into embeddings—think of them as a translator for large language models (LLMs). This “translator” infers context from a search query, letting a robust LLM sift through a vector database, validate the data, and decide if what it finds is relevant, correct, and valuable. Gone are the days of painstakingly structuring data for predictable retrieval.

Connecting the Dots: From Sales Engineering to raggity.ai

During my time as a sales engineer, I spent countless hours on activities that required deep domain expertise and an unwavering focus on customer experience. Beyond the demos, it was the follow-up emails and technical enablement sessions that consumed most of my day.? Every message needed to be as tailored as it was systematic—customized for each customer’s unique tech stack, whether that involved a specific cloud vendor, programming language, or set of tools.

I began to wonder: What if an AI could handle all these time-consuming, repetitive tasks? What if it could craft a follow-up email that’s as personalized as it is technically precise—freeing me to focus on advancing the deal?? That’s when the initial idea for raggity.ai was born.

TL;DR: All technical documentation of a product + all the technical know-how across every tech stack and language + an AI agent with specialized tools and domain expertise = raggity.ai.

Curious to see the early concept in action? Visit the raggity.ai site or watch the demo video below.

The Challenge: Crafting the Perfect Follow-Up

In my experience, follow-up emails aren’t just a quick message—they're the customer’s first step into our platform. They set the tone for the technical engagement that follows. Over the years, I’ve learned two key lessons from conversations with Sales and SE leaders, as well as buyers:

1. Any follow-up is better than none.
2. Buyers expect nothing less than a smooth, tailored experience.

Yet, when you glance at any sales engineer’s calendar, you'll see back-to-back demos and meetings. Often, the follow-up email ends up being nothing more than a list of links—a shallow gesture that barely scratches the surface of what the customer really needs. Each email should be customized for the customer's specific tech stack, highlighting common pitfalls and addressing the technical expertise of the audience. A bland, generic message not only confuses the customer but can also slow down the entire deal cycle. If the initial integration experience is clunky, it plants a seed of doubt: "If it’s this hard to set up now, what about long-term support?"

The Vision: Clarity Equals Speed

At its core, the perfect follow-up is all about clarity—because clarity drives speed. When teams evaluate multiple tools, the time it takes to integrate becomes a deciding factor. Every customer is looking for a way to cut down on extra work. For sales engineers—who simply cannot be experts in every combination of cloud vendor, language, and peripheral tool—the task of crafting such tailored messages repeatedly is nearly impossible.

A simple list of links won’t suffice. While a customer’s engineering team can certainly Google what they need, they might not know which products deserve their focus—especially when they’re exploring a new solution for the first time. They need a bit of hand-holding to get comfortable with the changes. An unclear or confusing message only leads to more questions, delays decisions, and can even tarnish the promise of ‘ease of use.’ It plants a seed of doubt: If installation is this challenging, how much time and support will the product demand in the long run?

In my view, achieving clarity is the key to speedy integration. This means delivering messages that are technically precise, acutely aware of the customer’s tech stack, and proactive in addressing foreseeable challenges. It’s a tall order—but an essential one for success.

That’s where an agentic RAG system comes in. (For clarity, “agentic” here means the system acts with autonomy and initiative, and RAG stands for Retrieval-Augmented Generation - topics I'll dive into in coming posts.) This system is designed to be the ultimate sidekick for sales engineers, delivering messages that are technically accurate, customer-specific, and proactive in addressing potential challenges. It transforms what once seemed like a tall order into a realistic, game-changing solution.

What’s Next?

In the following installments, I’ll dive into the technical nitty-gritty behind raggity.ai—exploring how I built it, the challenges encountered, and the lessons learned along the way. I’m excited to share the details and hope you’ll join me for the ride. Stay tuned!