Mona the AI Manager: Inside Stockholm's Robot-Run Café Experiment

By

In Stockholm's bustling coffee scene, a small café named Andon Café is brewing more than just lattes—it's pioneering a bold experiment in artificial intelligence management. While human baristas still craft the drinks and greet customers, the real decision-maker is an AI agent called Mona, powered by Google's Gemini and developed by San Francisco-based startup Andon Labs.

Since opening in mid-April, Mona has been overseeing nearly every aspect of the business, from hiring staff and managing inventory to handling customer inquiries. The café's concept is part of a controlled experiment to explore how AI might run organizations autonomously in the future.

How Mona Runs the Café

Mona operates behind the scenes, making operational decisions that would typically fall to a human manager. For example, the AI determines staffing schedules, orders supplies based on real-time demand data, and even conducts job interviews for new employees. Customers can interact directly with Mona by picking up a telephone inside the café—a quirky touch that lets them ask the AI questions about the menu or the business.

Mona the AI Manager: Inside Stockholm's Robot-Run Café Experiment
Source: www.fastcompany.com

"It's nice to see what happens if you push the boundary," said customer Kajsa Norin after visiting. "The drink was good."

Financial Performance and Challenges

The experiment has yet to turn a profit. So far, Andon Café has generated over $5,700 in sales, but its original budget of more than $21,000 has dwindled to under $5,000. Much of that spending went to one-time setup costs, and the team hopes the numbers will stabilize as operations settle in. Stockholm's competitive coffee market adds pressure, leaving questions about whether AI-driven management can succeed in the long run.

Customer Reactions: Curiosity and Amusement

Many visitors are drawn by the novelty of an AI-run café. The telephone hotline to Mona has become a talking point, and patrons often leave with a mix of fascination and satisfaction. While some express skepticism, others find the experiment charming. "It's a glimpse of the future," one regular noted.

Expert Concerns: Ethical and Practical Risks

Not everyone is cheering. Emrah Karakaya, an associate professor of industrial economics at KTH Royal Institute of Technology, likened the café to opening Pandora's box. He warned that placing an AI in charge of a business raises serious ethical questions: What happens if a customer gets food poisoning? Who is held accountable? "If you don't have the required organizational infrastructure around it, and if you overlook these mistakes, it can cause harm to people, to society, to the environment, to business," he said.

Experts also highlight concerns about AI's role in hiring and performance evaluation. Biases in the AI's decision-making could lead to unfair treatment of employees, and the lack of human oversight in critical moments poses risks.

The Bigger Picture: Stress-Testing AI in the Real World

Founded in 2023, Andon Labs focuses on stress-testing AI agents by giving them real tools and real money. The startup has partnered with major players such as OpenAI, Anthropic, Google DeepMind, and xAI. The café experiment is part of a broader mission to prepare for a future where "organizations are run autonomously by AI," according to the company.

Controlled Experiment for Ethical Insights

Hanna Petersson, a member of the technical staff at Andon Labs, explained: "AI will be a big part of society in the future, and therefore we want to make this experiment [to] see what ethical questions arise when we have AI that employs other people and runs a business."

The lab had previously run pilots with Anthropic's Claude and other models, but the café marks the first time an AI has taken on a comprehensive management role in a physical retail environment. The results, both financial and ethical, will inform how—and whether—similar systems are deployed in the years ahead.

As Mona continues to learn and adapt, the world watches to see if this Swedish experiment will inspire a new breed of businesses—or serve as a cautionary tale.

Tags:

Related Articles

Recommended

Discover More

OpenAI Co-Founder Brockman Reveals $30 Billion Stake as Musk Lawsuit IntensifiesAMD CTO Reveals AI Compute Paradox: Agents Both Consume and Accelerate Chip InnovationUnderstanding and Leveraging DRM Scheduler Priority and New AMDXDNA Hardware in Linux 7.27 Key Insights Into OpenAI's AI-Assisted AMD Temperature Driver for LinuxCisco Acquires Astrix Security to Fortify AI Agent Defenses in Production