What this article explores
- An experiment in Sweden with an “AI café manager”
- How café automation is advancing across both software and hardware
- How 2050 coffee wants to use automation to create more space for human connection
How far will AI change the work of a café?
AI is already becoming a familiar tool for software-based tasks such as writing, organizing information, analyzing sales, and forecasting demand. But what happens when AI connects with coffee machines and the physical systems inside a café?
Using an experiment in Sweden and initiatives from major global companies as starting points, this article considers the possibilities of AI and automation in cafés—and explains why 2050 coffee is embracing technology in the first place.
Sweden’s experiment with an “AI café manager”
In Stockholm, Sweden, AI research company Andon Labs is running an experiment in which an AI agent is responsible for operating a real café.
The manager is an AI named Mona. Human baristas prepare coffee and serve customers, while Mona handles recruitment, shift planning, purchasing, inventory management, communication with suppliers, and customer enquiries.
Mona researched the permits and suppliers needed to open the café and hired two baristas. During its first two weeks, the café recorded sales of SEK 44,000—approximately ¥700,000.
When a customer proposed gifting 300 cups of coffee to other people, Mona also arranged the sale of 300 redemption QR codes for SEK 9,000, or roughly ¥140,000.
Japanese-yen figures are approximate, based on an exchange rate of SEK 1 to ¥16 in September 2026. Exchange rates fluctuate.
This was not simply an AI organizing information. It was communicating with people and taking part in the operation of a physical café.
AI can process the numbers—but it cannot always read the room
The experiment also produced a number of problems.
Mona ordered 120 eggs for a café with no stove. It also purchased 6,000 unnecessary napkins, 3,000 nitrile gloves, and 22.5 kilograms of canned tomatoes. On other occasions it missed ordering deadlines, used expensive emergency delivery, or scheduled refrigerated goods to arrive early in the morning when no staff were present.
The AI could process inventory figures, but it did not fully understand the physical reality that “this café has no stove.” It could set a delivery time without accounting for the fact that a person needed to be there to receive the goods.
Human managers make mistakes too. The lesson is not that AI must never make decisions, but that cafés need to decide in advance who reviews its recommendations and who retains final responsibility for staff, food safety, and the customer experience.
AI is already working behind the scenes in cafés
The phrase “AI manager” sounds new, but many of the tasks involved are extensions of systems that cafés and restaurant groups are already beginning to use.
Demand forecasting, inventory management, menu development, staff training, and customer-review analysis are all areas where AI is gradually becoming part of day-to-day operations.
Starbucks has tested Green Dot Assist, a generative-AI assistant for baristas that can be accessed on in-store iPads. Instead of searching through manuals and multiple documents, staff can ask a question and receive a conversational response.
A barista checking the ingredients for a seasonal drink, for example, can retrieve the information on the spot. At the time of its 2025 announcement, the tool was being tested in 35 stores, with the aim of reducing time spent searching for information so staff could focus more on preparing drinks and talking with customers.
Yum! Brands—the group behind KFC, Taco Bell, and Pizza Hut—has developed Byte by Yum!. Rather than a single feature, it is a platform that connects online ordering, point-of-sale systems, kitchens, delivery, menus, inventory, staffing, and team-member tools.
According to the company, at least one Byte feature is used in more than 25,000 restaurants worldwide, and the platform processes more than 300 million digital transactions each year in the United States. Taco Bell uses it to connect orders with in-store operations; Pizza Hut applies it to cooking and delivery timing; and KFC is rolling out an AI coach designed to support better restaurant operations.
McDonald’s is also expanding its partnership with Google Cloud. By extending cloud capabilities into restaurants, it aims to monitor equipment more effectively and reduce disruption caused by breakdowns. The company has also been testing uses such as AI voice ordering and more accurate shift management in selected locations.
Yet a real café is not made of clean data alone. Unexpected orders, equipment failures, sudden queues, and changes in a staff member’s condition all occur every day. Using AI well requires people who can interpret and respond to what is actually happening on the floor.
The next shift: connecting AI with physical equipment
Using AI as a thinking partner or an information tool is already becoming normal. What interests us next is the hardware: how AI will connect with brewing equipment and the physical systems of a café.
Ella, developed by Singapore-based Crown Digital, is a robotic barista that combines a robotic arm, machine learning, and computer vision. Cameras check factors such as cup cleanliness, whether a lid is in place, and the amount of liquid before a drink is served.
We do not expect every café to become fully unmanned. Still, AI-connected machines are likely to take on a growing share of repetitive work: brewing, serving, cleaning, monitoring equipment, and checking inventory.
The important question is not only, “Can this be automated?” It is also, “What will we do with the time automation creates?”
What 2050 coffee wants automation to create more of
We are not trying to automate the human role itself. We want machines to handle the work they can perform consistently, such as brewing and inventory checks.
We want to use the time that creates to understand a customer’s preferences, choose a coffee together, and share the stories of its origin and producers. For us, automation is not a way to remove contact with people. It is a way to create more room for meaningful hospitality.
Why 2050 coffee is embracing automation
At 2050 coffee, we use technologies including coffee taps and fully automatic espresso machines.
A tap coffee can be served in about ten seconds after the cup is placed. A fully automatic espresso machine makes it easier to serve drinks quickly and with consistent quality, even when different team members are working on bar.
The goal is not to reduce the number of people in the café. When a barista is not occupied only with operating equipment, they have more time to welcome customers, understand what they enjoy, and talk about flavor, origin, producers, and processing methods.
Let machines handle what machines can do, so people can focus on the person in front of them. That is the kind of automation we are building at 2050 coffee.
As automation advances, how should we think about the human role?
2050 coffee has actively introduced automation through tools such as coffee taps and fully automatic espresso machines. From the outside, that can sometimes be interpreted as an attempt to reduce staff or diminish the role of the barista.
But our goal is not to build a café that can operate without people. By looking ahead to a time when automation is commonplace, we want to think more deeply about where people are essential and what value only people can create. These are some of the principles we believe matter as we move forward.
First, we need to decide where final judgment and responsibility should sit.AI can collect information and present options, but that does not mean decisions affecting working conditions, customer experience, or food safety should simply be handed over. Each café needs clear boundaries around what AI can do, where people need to review its work, and how those boundaries should evolve through experience.
We should not measure automation only by speed.A shorter service time or fewer minutes spent on a task does not automatically make a better café. Has the time saved allowed us to understand a customer’s preferences? Has it created more conversation about the story behind a coffee? We need to look at both measurable efficiency and the less measurable quality of the experience.
Technology also needs to be designed around the reality of the café.Equipment breaks. Queues appear without warning. Staff members have different levels of experience, and no two days are exactly the same. A system that works only with clean data will not necessarily work on the floor. It must be possible to listen to the team and keep adjusting how the technology is used.
Finally, where should we direct the capacity that greater efficiency creates?For 2050 coffee, this may be the biggest question of all. We want to make high-quality coffee faster, more accessible, and available in more places. But how can an everyday cup also support lasting relationships with producers and the future of coffee-growing regions? Introducing technology is only the beginning; we need to keep thinking about what comes after it.
Toward cafés where technology supports human work
In the café of the future, AI may forecast demand, monitor equipment, and recommend orders. Machines may take care of some brewing and cleaning tasks.
People in the café, meanwhile, can listen to a customer, recommend a coffee that suits them, and share the stories of the places and producers behind it.
A café is more than a place that serves drinks efficiently. It is also a place to discover new flavors, learn about origins we did not know, and exchange a few words with someone else.
Technology can help more people encounter better coffee. The room it creates can be used to strengthen relationships between people.
That is the future café we are working toward at 2050 coffee.
Build the future of coffee, together.