The Journey

Circa 2025
I spent the last several years building Scribble Data from the ground up—taking it from an idea in Indiranagar, Bangalore to a successful exit at York and Bay in Toronto. But that journey taught me something crucial: we were solving the right problem the wrong way.

At Scribble Data, we built an AI agents platform for Fortune 500 financial services companies. We worked on some of the gnarliest enterprise problems you can imagine — Group Benefits workflows, Pension Risk Transfer assessments, Forex forecasting, and FTP calculations in corporate banking. The kind of work where a single mistake costs millions and where "move fast and break things" gets you fired.

Here's what I learned: Enterprise AI adoption isn't struggling because the models aren't good enough. It's failing because we're trying to fit AI into human workflows instead of reimagining work around what machines do best.

Every day, I watched brilliant subject matter experts spend hours translating their knowledge into formats that AI could understand. I saw companies struggle with "AI stickiness"—getting initial wins but failing to achieve the NRR that had made SaaS so lucrative, back in the day. I saw eval frameworks disconnected from reality.

The problem wasn't the technology. It was what we were asking of it.

We kept asking: "How do we make AI work like humans?"

The right question is: "How do we redesign work so machines and humans each do what they're best at?"

Why Now, Why Me

I've been in the trenches of enterprise AI deployment. I know what kills projects in production—it's not model performance, it's organizational context, process integration, and the brutal reality of change management in Fortune 500 companies.

I've built the evaluation frameworks. I've integrated SMEs into AI workflows. I've seen what makes an AI implementation sticky versus what gets abandoned after the pilot. I've managed the pipeline, closed the deals, and lived through the pain of enterprise sales cycles.

But more importantly, I've experienced the exit. I know what it takes to build something valuable enough that others want to buy it. And I know that the real value wasn't in the technology we built—it was in understanding how work actually gets done in enterprises and then challenging those norms.

The Vision

My new company isn't about building better AI agents. It's about fundamentally rethinking how work gets structured when you put machines at the center of the design process.

Think about it: Every workflow in every enterprise was designed for humans. Job descriptions, org charts, approval processes, handoffs—all built around human constraints and capabilities. Now we're trying to bolt AI onto these human-centric workflows and wondering why adoption is hard.

We're building the infrastructure for machine-first organizations. Places where:

• Work is decomposed based on what machines excel at vs. what humans excel at
• Context flows seamlessly between AI agents and human experts
• The "work graph" is visible, measurable, and continuously optimizable

Why This Will Win

Timing: We're past the "AI is magic" phase. Enterprises have AI fatigue from failed pilots. They're ready for someone who actually understands deployment, governance, and production operations.

Unfair Advantages:

• I've deployed AI in environments where mistakes mean regulatory violations and eight-figure losses
• I understand both the technology stack (MCP, A2A protocols, context engineering) and the business reality (NRR, pipeline management, change management)
• I have the scars from building and exiting a company in this space—I know what worked and what didn't

Market Position: Every AI company is fighting to make their models smarter. We're enabling organizations to be fundamentally different. We're not a feature—we're a transformation.

The Ask

I'm not building another AI wrapper. I'm building the operating system for how the next generation of companies will work. Companies that are designed from the ground up for a world where AI isn't a tool—it's a colleague, a team member, a first-class citizen in how work gets done.

This is hard. It requires deep enterprise relationships, technical sophistication, and the patience to do true process re-engineering. But it's exactly the kind of hard that creates defensible, valuable companies.

I've already built and exited one company in this space. Now I'm building the one that solves the real problem, not just the obvious one.

.