Andrew Pignanelli wants to put entire companies on autopilot. Not just the boring parts—the whole thing.
His startup, The General Intelligence Company of New York, announced an $8.7 million seed round on December 8, 2025, led by Union Square Ventures. The pitch is audacious even by Silicon Valley standards: AI agent orchestration sophisticated enough that a solo founder could operate what looks, from the outside, like a full-fledged business. No engineers. No sales team. No support staff.
Just you, and a constellation of AI agents handling everything else.
Whether that's possible—or even advisable—remains an open question. But investors are intrigued enough to back it.
An Orchestration Layer for the AI Workforce
Pignanelli and co-founder Abhishyant Khare launched their flagship product, Cofounder, in September. The timing matters. They'd incorporated the company only months earlier, in January 2025, and moved quickly to market with a platform that organizes AI agents into something resembling corporate departments: engineering, sales, marketing, support, product, operations.
Each "department" coordinates multiple agents working on discrete tasks. An engineering agent might push code to GitHub. A support agent responds to customer inquiries. A marketing agent drafts email campaigns. The system integrates with services like Supabase, Vercel, and Postmark, stitching together a workflow that mimics, however imperfectly, what human teams do.
The platform includes human-in-the-loop controls for sensitive actions—a nod, perhaps, to the reality that fully autonomous business operations remain more aspiration than fact. Pricing starts at $20 monthly for the Pro tier, with usage-based charges stacking on top for agent activity, token consumption, and compute resources.
As of December 2025, the company claimed it had processed over 18 billion tokens per month and automated north of 59,000 tasks. Thousands of users signed up in the first week after launch, according to figures shared with Built In NYC. Those are the kinds of numbers that get venture capitalists paying attention, even if the long-term retention and utility remain to be seen.
The Backers
Union Square Ventures doesn't lead many seed rounds these days. That it chose to anchor this one suggests conviction in the broader thesis around AI agents, even if the one-person-company concept feels speculative.
Acrew Capital and Compound, both of which participated in General Intelligence's earlier $2 million pre-seed round announced in April, doubled down. They were joined by Agent Fund, The House Fund, and Untapped Capital. The angel roster includes Scott Belsky, Adobe's chief product officer and a veteran of the New York startup scene, alongside Yohei Nakajima, who came in through both Agent Fund and Untapped.
Additional checks came from Mute Ventures, basement.studio, and individual investors including John Phamous, Matthew Parkhurst, Sarah Chieng, and Jacky Huang. All told, General Intelligence has now raised $10.7 million.
Building in Public, Sort Of
The company announced Cofounder 1.5 alongside the seed round. By March, it had launched something called the "General Intelligence Fellowship"—$1,000 upfront plus $100 per day in platform credits for founders willing to test Cofounder by building actual companies with it. It's an interesting model: pay people to become your most demanding users.
In early May, beta access opened for Cofounder 2, which the company says expands its multi-department orchestration capabilities. The specifics remain somewhat vague, as they often do with early-stage AI infrastructure companies still figuring out what works.
Pignanelli aims to use the seed funding for enhancing agent infrastructure across key functions—expanding coverage across product development, engineering, go-to-market, support, and operations. In other words, making the vision more real.
The One-Person Unicorn Question
Here's the thing about the one-person, one-billion-dollar company: it's a seductive idea that brushes past some uncomfortable complexities. Running a business isn't just executing tasks. It's making judgment calls under uncertainty, navigating relationships, reading rooms, pivoting when something isn't working.
Can AI agents do that? Maybe eventually. But the founders building on Cofounder today are likely discovering what most entrepreneurs eventually learn—that the hardest parts of building a company aren't the parts you can automate.
Still, if Pignanelli and Khare can get even halfway there, it would represent a meaningful shift in how early-stage companies operate. Lower headcount, faster iteration, fewer dependencies. Perhaps that's enough to justify the bet Union Square Ventures is making.
The market will decide soon enough. Thousands of users signed up in the first week after launch in September 2025, already testing the premise.
