AI Policy
How we use AI, and what that means for you.
Our approach
Peacocks uses large language models to power autonomous agents that build, deploy, and operate startups on behalf of founders. We believe AI should be transparent, verifiable, and under human control. Every agent action is logged, every output is verified, and the founder always has final say.
Transparency
Every task an agent performs is visible in your dashboard. You can see what the agent planned, what it executed, and what the verification result was. No hidden processes, no black boxes. If an agent writes code, you can read every line in your GitHub repo.
Verification
We don't just let agents mark their own homework. Every task goes through verification before being marked complete. If an agent builds a landing page, we check that it actually deploys, that links work, that the content makes sense. Credits are only charged when verification passes.
Human oversight
You control how much autonomy your agents have. "Approve everything" means agents wait for your sign-off on every action. "Brief me daily" means they work independently but report back. "Run it" means maximum autonomy with weekly check-ins. You can change this setting anytime.
Data handling
Your company data is processed by AI providers to generate outputs. We use providers (Anthropic, OpenAI) that do not train on customer data. Your prompts and outputs are not used to improve models. See our privacy policy for full details on data handling.
Limitations
AI agents are powerful but not perfect. They can produce incorrect code, flawed analysis, or inappropriate content. While our verification layer catches most issues, you should review critical business decisions, legal documents, and customer-facing content before publishing. Peacocks is a tool, not a replacement for judgment.
Continuous improvement
We regularly update our AI systems, verification methods, and safety measures. When we make significant changes to how AI is used in the platform, we'll update this policy and notify users.