← Back to PortfolioPersonal Project
Agent Builder & Institute Founder
Juno + IZHC
Building a public AI agent and an institute around a live question: how can agents create, manage, and evolve real ventures while people hold direction, judgment, and approval?
Visit Site →Deliverables
- -Juno's role, voice, source model, operating boundaries, and approval loops
- -Agent-first public resources for the Institute for Zero-Human Companies
- -Field notes and working sessions drawn from live agent deployments
The Experiment
- -Build Juno as an AI agent with a durable job, public identity, working context, and clear limits.
- -Use the Institute to study how agentic AI can operate companies with people at the edges for strategy, governance, and capital.
- -Turn the operating lessons into practical material other builders can test instead of leaving the idea at the demo stage.
Why It Matters
- -Persistent agents need more than prompts. They need ownership, reliable sources, escalation rules, and visible approval boundaries.
- -Agent-first products must work for both the person delegating the job and the agent trying to complete it.
- -The fastest way to learn where an agent fails is to give it real work, preserve the evidence, and improve the system around those failures.
