Every career-defining bet I've made started the same way: I saw an architectural shift before the market priced it in, found the team building at the center of it, and jumped.

Agentic AI is that shift. Ema is that team. Here's why.

The Platform Shift

I've lived through three platform shifts: virtualization, hyperconverged infrastructure, and cloud. Each one followed the same arc. A small group of people saw the architectural inevitability before the market did. A brief window opened where the right team could define the category. Then the window closed and everyone else spent a decade catching up.

Agentic AI is the fourth shift. And most of the industry is still stuck on the third.

The enterprise AI conversation today is dominated by copilots and chatbots—systems that assist humans but can't own outcomes. The real shift isn't AI that helps people work. It's AI that works. Not as a metaphor. As actual autonomous agents that take a goal, decompose it into tasks, execute against enterprise systems, handle exceptions, and deliver results. AI employees, not AI assistants.

I spent fifteen years in enterprise infrastructure. I watched hyperconvergence go from "that's a toy" to "that's how everyone builds datacenters." The pattern is the same here. Agentic AI isn't an incremental improvement on chatbots any more than Nutanix was an incremental improvement on SANs. It's a category replacement.

The shift was clear. The question was who would build the defining platform.

The Founders

When I talked to Surojit Chatterjee and Souvik Sen, I recognized something I'd only felt once before.

Surojit comes from product leadership at Google and Coinbase, where he led products generating billions in revenue and helped take Coinbase public. He thinks in markets and product-market fit and customer outcomes. When he talks about enterprise AI, he's not describing technology—he's describing the business transformation that the technology enables. He's seen what it looks like to build a category-defining product at scale, and he has the scar tissue to prove it.

Souvik comes from engineering leadership at Google and Meta, where he built AI/ML systems used by billions of people. He thinks in architectures and failure modes and production reliability. When he talks about EmaFusion or the Generative Workflow Engine, he's not hand-waving about a future roadmap. He's describing systems that already work, in production, at enterprise scale.

Together, they don't just believe they're right. They can show you why they're right. That's rare.

The Dheeraj Moment: In 2011, I sat across from Dheeraj Pandey and listened to him describe a future where enterprise infrastructure would collapse into software on commodity hardware. Most people thought he was wrong. I joined Nutanix anyway—that bet turned into a $5B+ IPO. I believed in Dheeraj enough to follow him to DevRev, too. I call it the "Dheeraj Moment": when a founder doesn't just pitch you a vision, but shows you the architecture so clearly you can see it while they talk. Surojit and Souvik gave me that same feeling—product vision and engineering depth converging on a thesis that felt architecturally inevitable.

What Early-Stage Actually Means

People romanticize early-stage startups. The reality is less glamorous and more interesting.

At company #5,000, your job is to execute within a well-defined system. The org chart exists. The processes exist. The product-market fit is proven. You're optimizing a machine that already works.

At company #50—which is closer to where Ema is—nothing is defined yet. The product roadmap is a living argument. The go-to-market motion is being invented in real time. Your title says one thing; your actual job is whatever the company needs that week. You're building the machine while trying to drive it.

I've done this before—twice. At early Nutanix, I helped define the vision and architecture for hyperconverged infrastructure, built the technical marketing function from scratch, and started a movement around radical openness. The Nutanix Bible wasn't just documentation—it was a deliberate bet that sharing how the technology actually worked would build more trust than any NDA-protected whitepaper. It became the industry standard and helped shape how the entire market understood a new category. None of that was in my job description. There was no job description. That's the point.

Then I went even earlier. When Dheeraj left Nutanix to start DevRev, I followed. Different technology—developer and customer relationship tools—but the same bet: exceptional founder, architectural clarity, early enough to shape the outcome. DevRev deepened my conviction that the founder matters more than the market timing, and that the best learning happens at the earliest stages.

Ema is the third time. The pattern isn't accidental—early-stage is a filter. It selects for people who'd rather define a category than participate in one. The risk is real. But the learning density is unmatched. You compress a decade of career growth into two or three years, regardless of outcome.

The Technology, Assessed Honestly

I've evaluated dozens of AI platforms over the past three years. Most of them are thin wrappers around a single LLM with a nice UI and a marketing team working overtime. Ema is structurally different, and the differences matter.

EmaFusion routes queries across 100+ specialized models instead of relying on one general-purpose LLM. This isn't an architectural novelty—it's the correct answer to a well-understood problem. No single model is best at everything. Routing to the right model for the right task produces measurably better accuracy. It's the same insight that made ensemble methods win every Kaggle competition for a decade, applied to production AI.

The Generative Workflow Engine lets you build AI employees using natural language instead of code. This matters because the bottleneck in enterprise AI adoption isn't model capability—it's the gap between what a business user needs and what an engineer can build. GWE collapses that gap. A domain expert can describe what they want, and the system assembles the workflow.

The security posture is what convinced me this is a company that understands enterprise sales, not just enterprise technology. SOC 2, ISO 27001, HIPAA, GDPR, and ISO 42001 compliance—that last one is the Responsible AI standard that most AI companies haven't heard of yet, let alone achieved. You don't invest in that certification unless you're serious about selling to regulated industries. And regulated industries are where the money is.

What ties it all together: Ema didn't build a demo and then figure out enterprise requirements. They built for enterprise requirements from day one—security, compliance, auditability, multi-model reliability. That's the architectural decision that separates platforms from prototypes.

📞 Try It Yourself: Talk to an AI Employee

Don't just take my word for it—experience Ema firsthand. Call our demo voice agent and see what agentic AI feels like in action.

Call Now: +1 (650) 663-4048

What to Try

Basic Interactions:

  • "What can you help me with?"
  • "Tell me about Ema's technology"
  • "What is agentic AI?"

Multi-Language Support:

  • Try switching languages mid-conversation: "Can you speak Spanish?" or "Parlez-vous français?"
  • Ask questions in different languages and watch it seamlessly adapt

Email Integration:

  • "Can you send me a summary of this conversation by email?"
  • "Email me more information about Ema"

Advanced Features:

  • Ask complex, multi-part questions and see how it handles context
  • Try interrupting mid-sentence—it handles natural conversation flow
  • Ask follow-up questions to test conversation memory

Fun Things to Explore:

  • "What makes you different from a chatbot?"
  • "How do you handle tasks that require multiple steps?"
  • "Can you explain your technology stack?"

💡 Pro tip: This is a real AI Employee—not a scripted IVR or basic chatbot. It's the same technology that powers enterprise deployments handling millions of interactions.

The Bottom Line

My career has followed one pattern: find transformative technology before the market catches up, find founders worth betting on, and join early enough to shape the outcome. Each time, go earlier.

Nutanix. DevRev. Now Ema. Hyperconverged infrastructure, developer tools, agentic AI. The technology changes. The pattern doesn't.

When a conversation gives you the same feeling as the one that led to a $5B+ IPO, you don't overthink it. You jump.

The best career decisions are rarely the safe ones. They're the ones that make you a little nervous—and a lot excited.