"We almost over correct ourselves all the time because I came from a big company—one paranoia we have, healthy paranoia, is are we moving fast enough? And the answer almost always is no."
Vishal Sood pauses, considering the weight of what he just said. After 22 years in enterprise software, with 19 of those at Microsoft, he's now CPO and founding member of Typeface, an enterprise AI company that went from zero to a billion-dollar valuation in three years. The paranoia he describes isn't neurotic—it's strategic. In AI, standing still means falling behind.
The story begins in 2022, before ChatGPT existed in public consciousness. Abhay Parasnis, Adobe's former CPO and CTO, saw something others missed. Marketing teams were drowning in content demands but stuck with agency timelines and siloed workflows. "The way we do content is still kind of where data was 10 years ago," Sood explains. "We still operate in silos within a company."
Parasnis had a thesis: What if computers could finally handle creative work, not just number crunching? What if marketers could get personalized content whenever they needed it, tailored for specific audiences, channels, and geographies?
Today, Typeface does exactly that. The platform ingests brand guidelines, customer data, and marketing assets, learns what makes a company's voice distinctive, then generates content across email campaigns, social ads, and websites. When a marketer needs promotional emails for 100 different audience segments, Typeface creates them all, provides confidence scores, and feeds everything back into existing marketing systems.
For someone like Sood, the decision to abandon two decades at Microsoft wasn't about running from corporate life. It was about running toward something bigger.
"To me, the aha moment for AI was now it can do right brain things. So far we could do data crunching, data analysis, which is largely left brain problems." As a hobbyist photographer, he understood the significance. Creativity plus technology meant "a brand new chapter."
But the real revelation came after he made the jump:
"When it came to startup, all of you either succeed or you don't succeed. There is no question of four people on the table have different priorities. It just doesn't matter. And that is very refreshing. There is no convincing. There is no discussion. You just kind of go run, run, run."
In big companies, even when you're building something new, you're surrounded by people managing billion-dollar businesses. The conversation naturally gravitates toward optimization rather than breakthrough thinking. "You are kind of carving out your own thing, which is challenging," Sood reflects.
At a startup, everyone's fate is linked. Either the company succeeds or it doesn't. That alignment changes everything.
How does a company go from founding to billion-dollar valuation in three years? Part of the answer lies in who picks up the phone.
"We know a lot of people in big companies. So we have deep partnerships with Microsoft, Salesforce, Google now," Sood says. "That just helps you create distribution channels."
This isn't about getting deals through back-channel relationships. "You have to have a story, have to have a vision, you have to have the product," he clarifies. But when you need to get in front of enterprise decision-makers, having spent decades in their world opens doors that remain closed to unknown founders.
Sood has implemented something unusual at Typeface: If you can't ship something in three months, you probably shouldn't work on it at all.
"You wake up and things have changed around you. So if you're going to take six months to ship something or build something, you should really evaluate if somebody else is going to do it faster and better than you."
This philosophy extends beyond product development. Instead of writing traditional product specs, teams prototype directly using AI coding tools. "Writing a spec is harder than just opening up Cloud Code and just doing a quick proof of concept," Sood explains.
The result? Features that once took weeks now get built in days. Typeface has completely rebuilt its user interface three times in three years—not small improvements, but fundamental reimaginings. "It's a fundamental shift. It's not like a small incremental change here and there. And that's okay."
When Sood makes his pitch for why someone should join Typeface, he starts with scale: "The software industry all up is $650 billion, but the people industry is about $10 trillion. And that's the massive opportunity we have to make more efficient and disrupt."
The company's culture reflects this ambition through what he calls "authenticity"—being honest about what you're good at rather than chasing arbitrary growth metrics. Combined with what he describes as "ownership" and speed of decision-making, it's designed for the specific challenges of building enterprise AI.
There's something telling about Sood's journey. He didn't leave Microsoft because he was unhappy. He left because he recognized a once-in-a-generation technology shift and wanted to be at the center of it rather than on the periphery. Three years later, with Typeface valued at over a billion dollars, the bet appears to have paid off.
For professionals weighing similar decisions, his experience suggests that the combination of big company networks and startup urgency can accelerate both personal growth and company outcomes in ways that neither environment achieves alone.
For job seekers: Evaluate alignment ruthlessly. The biggest operational difference between big companies and startups isn't resources or processes—it's whether everyone truly shares the same definition of success or failure.
For recruiters: When hiring for enterprise startups, weight network effects heavily. The ability to open doors at target customers can compress sales cycles from years to months, making experienced hires worth premium compensation.
For founders: Institute aggressive shipping deadlines and mean them. Sood's three-month rule isn't arbitrary—in fast-moving markets, customer learning happens faster through working prototypes than through perfect specifications.