Zero Employees, One Human, a $250M Valuation — How Ben Cera’s ‘AI That Runs Whole Companies,’ Polsia, Even Automated Its Own $30M Raise
An AI company-operations platform built by ex-CloudKitchens operator Ben Cera with one human + AI. Nine agents run businesses 24/7; within months it reached ~$10M ARR (founder-reported run-rate) and ~7,600 businesses on the platform — and raised $30M at a $250M valuation with zero employees.
The pain point, and how they found it
The biggest reason indie builders and solo founders stall is simple: one person can’t run every function of a company at once — engineering, marketing, sales, support, finance. Hiring is faster but slow and expensive to do. Having run P&Ls and GMs across multiple markets at CloudKitchens, Ben Cera knew in his bones that most company operations are repeatable workflows — so they can be handed to agents. The pain wasn’t ‘a good idea,’ it was the labor to execute one, every day.
Polsia is a platform where you type in a business idea and a ‘team’ of AI agents runs the company for you. Nine specialized agents split the work — market research, landing pages, code, Meta/Google ads, cold email, customer support — sharing memory and running 24/7. At the center is an ‘Orchestrator’ (the CEO role) that drafts the day’s plan each morning and a summary each night. Pricing is outcome-aligned: $49/month plus 20% of the revenue your business earns, taken via Stripe.
The builder is Ben Cera. An operations person to the core — roughly 4.5 years as employee #2 at Travis Kalanick’s CloudKitchens, running P&Ls and GMs across multiple markets — he studied engineering at Columbia and now works out of Paris. He built Polsia solo, one human + AI, on about $1M of pre-seed he barely spent.
The numbers are loud. By his own account, within months of launch ARR reached a ~$10M run-rate, with around 7,600 business owners on the platform and 85% month-two retention. On speed, he also claims ‘$1M ARR in 30 days’ and ‘$6.27M ARR in under 90 days.’ Then, in May 2026, Polsia raised $30M at a $250M valuation from a syndicate including Sound Ventures and True Ventures — still with zero employees. The set piece: the AI ran the fundraise itself. Polsia assembled the data room, handled investor briefings and diligence, and Ben says he only showed up for the final calls. The product demo was the company’s own financing round.
But the figures invite skepticism. ARR here isn’t realized revenue — it’s a founder-reported run-rate (and it folds in an aggregate of customers’ revenue × 20%), and whether early customers retain at scale is unproven. The next section breaks down the nine-agent design, the outcome-based pricing, and the ‘make the raise the demo’ strategy — separating what’s reproducible from what’s narrative.
The repeatable playbook
- 1Decompose company operations into repeatable workflows and assign an agent per function (CEO/eng/marketing/sales/support/finance…)
- 2Give agents shared memory (context) and run one autonomous cycle each night (nightly task + on-demand credits)
- 3Make pricing outcome-based ($49/mo + 20% of revenue via Stripe) — bet on customer success to win trust and price ceiling at once
- 4Demo the product on your own operations/raise (let AI run the data room and investor process; keep a public dashboard for transparency)
- 5Build in public relentlessly on X + launch on Product Hunt for initial momentum (a bold story is the biggest acquisition engine)
- 6Talk about ARR by distinguishing realized revenue from run-rate, and answer skepticism head-on (to earn trust)
Polsia’s numbers invite legitimate skepticism. ARR isn’t realized revenue — it’s a founder-reported run-rate that folds in an aggregate of customers’ revenue × 20%, so if early customers don’t retain, the headline ARR inflates without substance behind it. The $250M valuation has been called frothy, and Mixergy confronted him with ‘is this a scam?’ Yet real signals also exist: ~7,600 businesses on the platform, 85% month-two retention, brand-name VCs. The conclusion isn’t binary — discount the figures as a reached run-rate, and extract only the reproducible mechanics: agent division-of-labor, outcome-aligned pricing, and making the raise itself a demo.
Deep dive
【Deep dive】What makes Polsia interesting is that it ‘productized’ the fuzzy job of running a company through three levers: agent division-of-labor, outcome-based pricing, and narrative. We dissect each — separating what’s reproducible from what deserves skepticism.
■ Why ‘running a company’ can be agent-ified (the nine-agent split). Polsia’s core is decomposing company work by function and assigning each to a specialist agent. Per reporting and the founder, the lineup is roughly: an Orchestrator (CEO — morning plan, evening summary), social (drafts and posts), email outreach (finds prospects, sends cold email), support (reads the inbox, drafts replies), ads (optimizes Google/Meta), finance (syncs Stripe revenue, tracks spend), business planning (updates strategy and KPIs), competitor research (web searches), and code generation (ships features, opens PRs). The key is shared memory: every agent sees the same context, so fewer things slip. Ben Cera’s CloudKitchens insight — most company operations are repeatable workflows — is the architecture itself. The takeaway for indie builders: split your own recurring operations by function and run them on agents plus shared memory.
■ Outcome-based pricing ($49 + 20% of revenue) — clever and double-edged. Pricing is a thin $49/month fixed fee plus 20% of the revenue the business generates. It works two ways: (1) betting on the customer’s success earns trust, and the price ceiling rises with the customer’s growth; (2) because Polsia earns more when customers earn more, the agents are motivated to maximize revenue. But it’s double-edged: the headline ARR folds in an aggregate of ‘customer revenue × 20%,’ so if customers’ revenue is transient, Polsia’s ‘ARR’ inflates on a run-rate basis while realized, retained revenue is a separate question. Read the number as a run-rate, not realized MRR.
■ Turning the fundraise itself into the ultimate demo. Polsia’s biggest ad is the fact that the AI ran the $30M raise. Polsia assembled the data room, ran investor briefings and diligence; Ben only joined the final calls. Two effects: first, it’s the ultimate case study — proving the product on a live, high-stakes job (your own financing). Second, combined with build-in-public on X and a public dashboard, the funding news became a giant PR engine. Sound Ventures, True Ventures and others participated as a syndicate (no single lead was named). The growth core wasn’t ad spend — it was a hard-to-rebut story: ‘$250M with zero employees.’
■ The economics of the solo founder (80% AI, 20% taste). Zero employees means erasing payroll, the dominant fixed cost. Gross margin on the SaaS portion approaches ~100% in theory, and real costs shift to LLM inference and ads. Ben describes the work as ‘80% AI, 20% taste’: hand implementation and operations to AI; the human concentrates on direction, judgment, and final accountability. It’s a claim that rewrites what ‘owning a company alone’ costs — and if you want to copy it, the crux is deciding what to automate and which 20% you keep.
■ Skepticism and the real debate (don’t skip this). Polsia draws persistent doubt. Mixergy grilled him with ‘is this a $250M scam?’, and multiple reviews flag run-rate inflation, unproven retention of early customers, and a frothy valuation. Fairly read: the strong signals are real (~7,600 businesses on the platform, 85% month-two retention, brand-name VCs). The weak points are equally clear (ARR is a run-rate, not realized, and includes an aggregate of customer revenue; churn resistance needs time; ‘AI ran the raise’ is partly narrative). The practical conclusion for indie builders: discount the figures as a ‘reached run-rate,’ and copy not the headline valuation but three things — (a) agent-ifying recurring operations, (b) outcome-aligned pricing, and (c) making the launch/raise itself the acquisition engine.
■ Takeaways for indie builders (summary). First, ‘I can’t scale because I can’t hire’ is no longer an excuse — recurring work runs on agents plus shared memory. Second, aligning price with outcomes wins trust and price ceiling at once (just read ARR carefully). Third, the biggest marketing isn’t flashy ads; it’s showing your own operations or fundraise as a live demo, transparently, on X. Polsia went after the win not with ‘smarter AI’ but with productized company-operations plus deliberate narrative design.