How General Intelligence used agents to build an agent platform on Vercel
Source: Vercel Blog
General Intelligence on Vercel
General Intelligence is building a platform that lets any founder run a company entirely with AI agents. Their vision is a one‑person, billion‑dollar company where every department is driven by agents.
Their flagship product, Cofounder, gives founders a full team of agents covering engineering, marketing, SEO, finance, sales, customer support, and operations. The company is an 8‑person team with 5 engineers. To ship a platform that lets customers run agentic companies, they had to operate as one themselves, using Cofounder’s CTO agent to build other agentic business functions. They quickly discovered that the challenge was infrastructure, not the agents themselves.
Why Vercel?
For a complex, multi‑tenant platform the requirement was total programmatic control: every action a human could take on the underlying cloud platform had to be available to a coding agent through a CLI or APIs (e.g., killing deployments, changing DNS, managing billing, editing configs). Most cloud providers fell short, so General Intelligence migrated to Vercel.
- Vercel provides a single CLI, a unified API, and a consistent observability layer.
- Every change the CTO agent makes goes straight to a Git branch, spawns a preview environment, and is tested end‑to‑end by a browser agent on the live URL.
- The team now has over 4,000 preview branches in flight, with roughly 100 versions of the app running concurrently, each driven by a coding or browser agent.
Impact on Engineering
- Engineers ship an average of 10 PRs per day, each with 70+ commits.
- Token budget: $5,000 per engineer per month.
- The Python backend was one of the first complex stacks migrated to Vercel, allowing the agent to drive deployments, configs, and compute across the whole stack—not just the frontend.
Customer Experience
When a founder spins up a company on Cofounder they receive:
- A real GitHub repository.
- A managed Vercel deployment, provisioned automatically.
- Their own domain with SSL and DNS handled automatically.
The same CTO agent runs inside each customer’s company, following the same workflow of branches, preview environments, and browser‑based testing.
Agents Need Infrastructure Built for Agents
Traditional cloud platforms are designed for human developers: dashboards to click, consoles to read, and APIs that cover only a subset of operations. A coding agent runs dozens of processes in parallel, treating logs as data, parsing errors as input, and interpreting every dashboard click as a missing API. The bottleneck becomes how much of the cloud the agent can actually reach, not how fast it can write code.
Vercel’s model—full programmatic control via CLI and APIs—eliminates those gaps, enabling agents to perform 100 % of the required work.
How Five Engineers Ship Like a Hundred
- Automation: 90 % of SRE work is automated through Vercel and Cofounder’s own agents.
- Preview‑First Workflow: Every change creates a preview branch, instantly testable on a live URL.
- Unified Tooling: One CLI, one API, one observability layer reduces surface area for both developers and agents.
- Scale: 4,000+ preview branches and ~100 concurrent app versions provide massive parallelism without additional human effort.
What’s Next
General Intelligence plans to continue expanding the platform, adding new departments and agents while leveraging Vercel’s ability to let them focus on customer value rather than cloud configuration. The ultimate goal remains a fully autonomous, AI‑driven company that any founder can launch with a single click.