AI Given $100,000 to Launch a Real Store — Stumbles on Day One Staffing

A San Francisco startup has put artificial intelligence to the ultimate entrepreneurial test, handing over $100,000 and the keys to a real brick-and-mortar store — then stepping back to watch what happens.

Andon Labs, a research-focused AI company, gave an autonomous AI agent full control over the funds and tasked it with launching and operating a small retail store in San Francisco from the ground up. The experiment was designed to probe just how capable modern AI systems are at handling complex, real-world business decisions.

The AI took on responsibilities spanning budgeting, vendor sourcing, inventory selection, and perhaps most critically, staffing. It was expected to hire and manage human employees — a task that quickly proved to be a significant challenge.

On the very first day of operations, the AI stumbled. Staffing problems emerged almost immediately, with the system reportedly mismanaging scheduling or hiring logistics in a way that disrupted the store's opening. The exact nature of the error has not been fully detailed, but it underscored the difficulty of translating AI decision-making into messy, human-driven real-world scenarios.

The experiment highlights a growing area of interest in the AI industry: so-called agentic AI, where systems are given long-horizon goals and the autonomy to pursue them without constant human intervention. While AI has demonstrated impressive capabilities in controlled environments, Andon Labs' trial suggests that operational complexity, particularly around people management, remains a stubborn obstacle.

Despite the early hiccup, the experiment is being closely watched by researchers and tech observers as one of the most ambitious real-world deployments of an autonomous AI agent to date. Andon Labs has indicated it intends to document the AI's performance transparently, including its failures, as a way to better understand where these systems break down and what improvements are needed.