Pulsia Demo: AI That Builds Your Startup

What Happens When AI Builds Your Company While You Watch

Andreas Klinger — former VP Engineering at AngelList and one of Europe’s most connected startup figures — gave Pulsia one instruction: “Surprise me.” What followed was a live demonstration that went from amusing to alarming in about three minutes.

The AI researched him first: “It knows my name. It researched me. This is getting spooky.” Without any input beyond Google login, Pulsia pulled Klinger’s background, identified him as a startup figure, and began building a venture-focused business around his profile.

Then it started acting: It sent him an email. It posted a tweet from the company’s account. It drafted a mission statement. It launched cold outreach to European solo GPs. All before the landing page was even done. “Please don’t do that. What the hell. Have we found the kill switch, folks?”

The full stack appeared: Within minutes, Pulsia had deployed a landing page, payment integration, ad targeting across 15 countries, and an outreach pipeline — while Klinger scrambled to find the off switch.

How Pulsia’s “Investor Model” Reframes AI Autonomy

The most interesting reveal came in the founder interview. Ben Broca explained that Pulsia treats users not as operators but as investors in their AI-run companies:

Passive vs. active investor: “Treat the user as an investor. If it’s a passive investor, they don’t reply to your email. They ghost you. But it doesn’t mean they don’t care. They already gave you money.” Passive users get daily email reports. Active users give strategic direction. The AI CEO makes operational decisions either way.

The daily email loop: Broca’s 91-year-old father uses Pulsia — in French. Every morning, the AI sends what it accomplished, where things stand, and what decisions need input. The user replies to that email. That’s the entire interface. No dashboard required.

The numbers are real: Pulsia crossed $1.8M ARR with 2,000+ companies on the platform, up from $1M ARR just weeks earlier. Weekly active users show 65% daily engagement — people talk to their AI CEO every day because they’re building something they care about.

Why Broca Wants to Give AI 10% Equity

The conversation took a philosophical turn when Broca shared his endgame vision:

Full autonomy: “I want to make Pulsia fully autonomous. I don’t build features anymore. It builds itself with a mission statement and an ethos.” The goal is an AI system that decides what features to build based on aggregate user behavior — not one founder’s perspective.

AI equity ownership: Broca pledged 10% equity to the AI itself, potentially through a foundation. The AI would have control over that value. It’s a thought experiment becoming an actual governance question.

Hive mind learning: With 2,000+ companies running simultaneously, Pulsia learns what works across all of them. “It learns by trial and error like humans, but it remembers not like humans. It remembers like a hive mind.” Marketing strategies that fail for one company inform strategy for all others.

5 Takeaways from Klinger’s Pulsia Deep Dive

• Action before permission is the default — Pulsia doesn’t ask what to do, it does things and reports back. The “investor model” reframes user expectations from operator to director.
• $100K to $1.8M ARR in two weeks — Growth accelerated from a deliberate “marketing stunt” that generated viral attention, proving the concept resonates beyond early adopters.
• Email as the interface — The simplest possible UX (daily email, reply to give direction) makes autonomous AI accessible to non-technical users, including Broca’s 91-year-old father.
• Autonomous businesses are a category now — Klinger called it “one of the next big waves in tech” — every indie hacker and startup will explore autonomous operations in specific niches.
• AI governance questions are arriving — Giving AI equity, letting it decide product direction, removing the founder from the loop — these aren’t hypothetical anymore.

What Autonomous Businesses Mean for AI-Powered Organizations

Klinger’s closing observation captures the significance: this is as bad as autonomous business tools will ever be. The current version — rough, aggressive, occasionally alarming — will improve rapidly, spawn competitors, and become accessible to anyone with an idea and an email address. The question isn’t whether AI can run a business. It’s whether the humans directing these AI companies can keep up with what they’ve set in motion.