A Solo Founder's AI Agents Build His Product, Run His Ads, and Handle Support. He's at $800K Run Rate.

Ben Cera has zero employees. His company, Polsia, hit an $800,000 annual run rate less than two months after launch. The product was built by AI coding agents. The ads are created by AI. The customer support is handled by AI. The bugs are fixed by AI.

This is not a thought experiment about the future of work. This is a company generating real revenue right now with a single human making decisions and a fleet of specialized agents doing everything else.

Andrew Warner interviewed Cera on Mixergy this week. The conversation is worth listening to in full, but the core revelation is this: we have crossed a threshold where a single person can operate a full company, product development, marketing, sales, and support, using nothing but AI agents.

And the implications for anyone thinking about starting a business in 2026 are significant.

How Polsia actually works

Polsia is a platform where a user brings a business idea and AI agents set up everything needed to run it. The architecture breaks down into specialized agents, each with access to the tools they need.

An engineering agent has access to a web server, a GitHub account, and a database. It writes code, deploys it, and fixes bugs when they appear. A marketing agent has access to a X account, an email address for cold outreach, and a Meta Ads account. It creates campaigns, generates ad creative, and manages spend. A support agent has access to an email inbox and handles customer inquiries.

When a new user signs up, the system automatically provisions everything: an email address, a virtual server, a database, a Meta advertising account, an OpenAI API key for video generation. The user describes their business idea. The agents build it.

The UGC video ads that AI makes

One of Cera's examples was an HVAC company. The AI generated a video of a man who appeared to be talking into his phone camera, saying he used to hate dealing with marketing agencies and now the tool handles everything so he can just focus on fixing AC units. The person in the video was not real. Sora generated it.

The ad linked to a landing page that the engineering agent had built. The support agent handled inbound questions. The entire customer journey, from ad impression to signup to post-purchase support, was operated by agents.

This is a qualitative shift from "AI helps with marketing" to "AI is the marketing department, the engineering team, and the support staff."

Why this matters for you

The specific details of Polsia are interesting, but the underlying pattern is what matters. A single person can now deploy specialized AI agents that cover the core functions of a small business:

Product development. Coding agents build features, fix bugs, deploy updates. They are not perfect. They need oversight. But they move fast enough that one person reviewing their output can ship at the pace of a small team.

Marketing. Agents that can generate video creative, write copy, set up ad campaigns, and manage budgets autonomously. The cost of customer acquisition drops because you are not paying an agency or a marketing team. You are paying for compute and ad spend.

Sales and support. Agents that respond to inbound inquiries, qualify leads, and handle post-purchase questions. They operate around the clock. They do not need PTO.

The total cost of running these agents is measured in hundreds of dollars per month, not tens of thousands. That changes the minimum viable economics of starting a company.

Five ways to build on this pattern

Build a vertical AI-run business

Polsia is horizontal. It builds whatever the user describes. The bigger opportunity might be vertical. Pick a specific industry, plumbing, landscaping, personal training, real estate, and build the entire AI-run business specifically for that niche.

The AI agents know the industry terminology, the common customer objections, the seasonal patterns, and the typical service offerings. A plumber signs up, the agents build a website optimized for local search, create video ads featuring AI-generated testimonials specific to plumbing, set up automated booking and follow-up, and handle customer communications.

Price it at $200 to $500 per month. A plumber paying $300 per month for a full marketing and operations system is getting a fraction of what they would pay a marketing agency. At 500 customers, that is $150,000 in monthly recurring revenue.

AI Business-in-a-Box consulting

Not everyone wants to use a platform. Some founders want a custom setup. Offer a service where you configure the full agent stack for their specific business: coding agents pointed at their codebase, marketing agents connected to their ad accounts, support agents trained on their product.

Charge $5,000 to $15,000 for the initial setup. Add a $500 to $1,500 monthly retainer for monitoring and optimization. You are not writing code for them. You are configuring and orchestrating agents.

The skill you are selling is knowing which agents to use, how to connect them, and how to set up guardrails so they do not go off the rails. That is a real skill, and it is rare right now.

Autonomous ad creative generation

The piece of Polsia's stack that is most immediately monetizable on its own is the ad creation pipeline. An agent that takes a product description, generates a UGC-style video ad with an AI-generated person, writes compelling copy, and uploads it as a Facebook campaign is something every small business would pay for.

Build this as a standalone service. Price it at $99 to $299 per month, which includes a set number of ad creatives per month and automated campaign management. The cost to generate each video with Sora is under $1. The Meta API is free. Your margin is enormous.

Target local businesses and e-commerce brands that are currently paying $2,000 to $5,000 per month to agencies for the same output. Your service delivers faster, costs less, and scales without adding headcount.

AI agent management retainer

Justin Parnell, another solopreneur profiled by Business Insider this week, quit his VP of marketing role at 36 to build custom AI agents for other entrepreneurs. His agents handle everything from lead scoring to proposal generation to invoicing.

"When someone fills out a form on my website, they receive a custom road map for building agents. Another agent generates a proposal and sends it along with a meeting link to book time on my calendar."

He charges clients for agent setup and ongoing management. The monthly retainer model works because agents need tuning. Prompts drift. APIs change. New use cases emerge. Someone who understands the agent stack and can keep it running smoothly is worth paying.

Charge $1,000 to $3,000 per month for managing a client's suite of AI agents. Five clients at $2,000 each is $10,000 per month.

A course on building AI-run companies

The demand for this knowledge is growing faster than the supply. Business Insider is running multiple profiles of solopreneurs using AI agents. Entrepreneur Magazine published two articles about AI tools for solo businesses in the same week. The topic is hitting mainstream awareness.

A course that teaches non-technical founders how to set up their first coding agent, connect a marketing agent to their Meta Ads account, build a support agent that handles email, and orchestrate all three into a functioning business would have strong demand.

Price it at $300 to $700 per seat. Run cohorts of 20 to 30 people. The first two cohorts are your case studies for marketing the rest.

The economics

Let us be specific about what this costs to run. Sora video generation runs about $0.50 to $1.00 per video at current API pricing. Claude or GPT coding agent usage for a typical development session costs $5 to $20. Meta Ads API access is free. Email and support infrastructure is $20 to $50 per month.

A fully AI-operated small business can run its agent stack for $200 to $500 per month in compute costs. Compare that to the $8,000 to $15,000 per month it costs to employ even a small team of a developer, a marketer, and a support person.

That is a 95% reduction in operating costs. The quality is not equivalent to a senior team. But for a small business, it is more than sufficient to compete.

What to take from this

Two months, zero employees, $800,000 in annual run rate. That is not a projection. It is Polsia's actual numbers as of March 2026. The AI agent stack is production-ready for running real businesses. The question is not whether this will happen. It is happening. The question is whether you will build one of these businesses or help others build theirs.