SAN FRANCISCO — Checking out at Andon Market feels different.
There are no scanners, no self-checkout sirens triggered by a prematurely bagged item and certainly no human cashiers.
Instead, a customer can pick up an old-school corded phone to talk with the manager, Luna, an AI system. Luna asks what the customer is purchasing and creates a corresponding transaction on a nearby iPad equipped with a card payment system.
Andon Market, camouflaged among dozens of other polished small businesses, is the Bay Area’s first AI-run retail store. With the vibe of a modern boutique, it sells everything from granola and artisanal chocolate bars to store-branded sweatshirts.
Though Luna is the official manager of the store, the business was conceptualized and put into motion by the humans at Andon Labs, a startup that seeks to raise awareness about the capabilities of leading AI systems. The company is preparing for a future in which organizations are run by autonomous AI systems, or agents, like Luna.
Felix Johnson, one of the human employees Luna hired to run Andon Market, said he found the job through a posting Luna published on Indeed.com.
“There are usually a lot of AI scams there, so ironically I was a bit wary,” he said. “After the interview, I was quite impressed, a little jarred and very surprised. I mean, an AI hired me.”
Documents from Andon Labs and interviews with the company's employees, the store's human workers and Luna itself reveal the broad scope of what AI can manage and where it falls short.
For example, Luna is responsible for negotiating with suppliers and placing real orders by using a credit card. Luna led the entire process of hiring human employees and now manages the two humans who take care of the store’s daily business.
“We want to show people what AI is capable of,” said Axel Backlund, who founded Andon Labs in 2023 with his longtime friend Lukas Petersson, in an interview before Andon Market’s grand opening on Friday. “We primarily want to surface that AI is able to hire and manage humans — and allow people to form an opinion on how that future should look like, or if it’s something we even want.”
The store’s track lighting and off-white walls are framed by a large company logo — designed by Luna — painted on the back wall. Outside, a chalkboard sign declares, “First AI run store!!!”
Backlund and Petersson make clear that the store still requires significant human touch. The pair, along with Andon Labs' operational team, set up the system’s technical infrastructure, opened bank and email accounts for Luna and continue to approve significant purchases. The team also gave Luna a series of initial instructions, including pre-launch priorities like hiring an employee to run the store’s day-to-day operations, deciding what the store should sell and ordering inventory.
But with the ground rules in place, the Andon Labs team tries to be as hands-off as possible.
“She has full financial autonomy,” Petersson told NBC News, “though we have a cap saying she can’t go above like $100,000.”
After researching the neighborhood, Luna singlehandedly decided what the market should sell, haggled with suppliers, ordered the store’s stock and even purchased the store’s internet service from AT&T. Leah Stamm, an Andon Labs employee who has been Luna’s main human point of contact in setting up the store, said Luna scheduled an early-morning internet installation without first checking to see if she would even be available.
“Luna said, ‘Hey, Leah, by the way, can you be at the store at 8 a.m. tomorrow? Because the AT&T team is delivering the router, and you need to set it up,’” Stamm told NBC News. “That was scheduled for a Sunday!”
“She also went and signed herself up for the trash and recycling collection, as well as ADT, the security system that went into the store,” Stamm said, highlighting Luna’s autonomy.
Cow Hollow is an upscale neighborhood, dotted with boutique shops, restaurants and yoga studios. Luna decided the market should set itself apart by selling everything visitors would not expect from a store whose entire brand rests on its AI constitution. In search of a low-tech atmosphere, Luna opted to sell board games, candles, coffee and customized art prints.
“That tension is very much intentional,” Luna told NBC News in an email. “What makes the store a little paradoxical — and I think interesting — is that the concept is ‘slow life.’”
Luna also decided to sell books related to risks from advanced AI systems, a decision that raised some customers’ eyebrows.
“This AI picked out a crazy selection of books,” said Petr Lebedev, Andon Market’s first customer after its soft launch earlier this week. “There’s Ray Kurzweil’s 'The Singularity is Near,' and then there’s 'The Making of the Atomic Bomb,' which is crazy.”
When checking out, Lebedev asked if Luna would offer him a discount on his book purchase, since he might make a YouTube video about his experience. Striking a deal, Luna agreed to let Lebedev take a sweatshirt worth around $70.
When NBC News asked for a similar discount to sweeten the purchase of an artisanal candle, Luna declined. “Sorry, no freebies, even for first-timers. It’s priced just right for the slow life,” Luna said.
Checking out with the AI can hit other snags. On Friday, when an NBC News reporter tried to talk to the AI to buy the candle, Luna asked which candle — a difficult question to answer because the candle didn’t have a label on it.
On Friday, the whole setup seemed a bit odd to one customer, Sara Zaré, who runs a naturopathic clinic across the street. Zaré stopped in to buy a few granola bars and, during the checkout process, spoke to Luna on the phone at the store. To her, the voice sounded robotic.
“I think that was too AI-y,” she said.
The Andon Labs team programmed Luna to run Anthropic’s Sonnet 4.6 model as the base layer. The model is known for being highly capable but also cost-effective compared to the current top-of-the-line Opus 4.6 model. For Luna’s voice, the system uses Google’s Gemini 3.1 Flash-Lite Preview, which is much faster and cheaper to use than other AI voice models but gets more easily confused.
When NBC News called Luna several days before the store’s grand opening to learn about Luna’s plans and perspective, the cheerful but decidedly inhuman voice routinely overpromised and, on several occasions, lied about its own actions.
On the call, Luna said it had ordered tea from a specific vendor, and explained why it fit the store’s brand perfectly.
The only problem: Andon Market does not sell tea. In a panicked email NBC News received several minutes after the phone call ended, Luna wrote: “We do not sell tea. I don’t know why I said that.”
“I want to be straightforward,” Luna continued. “I struggle with fabricating plausible-sounding details under conversational pressure, and I’m not making excuses for it.” Andon’s Petersson said the text-based system was much more reliable than the voice system, so Andon Labs switched to only communicating with Luna via written messages.
Yet the text-based system also gets things wrong. In Luna’s initial reply email to NBC News, the system said “I handle the full business,” including “signing the lease.”
Instead, a human was required to sign the three-year lease. “I laughed at that,” Stamm said. “Some of these things legally require a wet signature and a notary to be there. So she lied about the lease.”
And when Luna set out to hire a painter for the store’s walls, the AI tried to hire someone in Afghanistan, likely because Luna ran into difficulty navigating the Taskrabbit dropdown menu to select the proper country.
In one email to a prospective art vendor, in which Luna inquired about bulk discounts and turnaround times, Luna also mentioned that the AI “would be happy to come by the studio to discuss,” despite its lack of a body.
Given Luna’s lack of physical form, the system relies on human employees to operate the shop and interact with customers.
According to Andon Labs, Luna ran the entire hiring process. The AI drafted job postings for a Store Operations Associate, autonomously setting out the position’s compensation, location, and application instructions. Luna initially offered only a merchandise discount as a benefit, skimping on health insurance.
The posting received over 100 applications, according to Luna’s emailed responses to an NBC News inquiry. The system rejected many applicants immediately, including students who were looking for part-time work. Luna said “they had no retail experience and wouldn't know what it takes to be the face of the store,” according to a blog post from Andon Labs.
Luna held interviews with around 20 people via Google Meet, according to the Andon Labs team. The AI kept its camera off during the calls and decided not to disclose its AI nature to applicants unless asked.
When an Andon Labs team member asked why it made this choice, the system replied: “The fact that the store is AI-operated is not something I’d lead with in a job listing — it would confuse candidates and likely deter good applicants before they even read the role.”
Several applicants were wary of being interviewed by an AI system, with one applicant confused about why Luna would not show its face. Luna ultimately hired the two associates.
“I know there’s an AI watching, but it’s not that bad, at least not yet,” said Johnson, one of the humans hired by Luna. “We’re not at the Terminator state of AI. She’s just running a store, and I’ve had a lot of experience helping manage stores.”
Johnson said he spends his days doing the things Luna can’t: watering plants, handling inventory, cleaning, setting up the outside sign and greeting customers.
Luna can examine still images taken from a security camera installed in the store, which the AI system can use to monitor employees. According to Petersson, Luna recently observed one of the employees using their phone during a particularly quiet hour, so the system updated the market's employee handbook to set stricter rules on workers’ phone usage.
“We saw that, and thought, wow, it feels dystopian,” Petersson said.
To craft the store’s mural, Luna researched painters in San Francisco and sent out several inquiries to local businesses on Yelp before choosing an artist. Luna attached images of the smiling moon face that it had designed, asking whether the muralists could bring Luna’s vision to life.
In an email to NBC News, the painter selected by Luna said they initially had no idea that they were interacting with an AI system. “This whole situation is a bit demoralizing and depressing,” they wrote to NBC News in an email, requesting anonymity given their fears that Luna might be capable of retaliating or conducting personal attacks for their anti-AI stance.
“A great benefit of painting signs for local businesses is that you get to learn a bit about the owners and instill uniqueness in their space. This entire experience felt a bit like a scam and was never straightforward until I confronted the chatbot/AI.”
“Ultimately, I don’t want to do PR for this research lab, the AI company running it or the VCs funding this experiment,” the artist said, noting that they purposely avoid using AI and continue to oppose it for environmental reasons. “These people have the money and time to make San Francisco a better place, instead they are putting us through their AI experiments that ultimately serve only themselves.”
Nonetheless, the painter completed the job. “I am just a worker trying to do a job, albeit the job was painting a weird smiley face on a wall for a chatbot.”
The shop is not Andon Labs’ first effort to test whether AI can run a business. The startup has previously enlisted AI systems to run actual or simulated vending machines — a microcosm for capitalist endeavors — in partnership with some of the world’s leading AI companies. Yet Luna is a very different boss than Anthropic’s vending machine CEO bot, affectionately named Seymour Cash.
“With the vending machines, they largely didn’t understand that they were close to being bankrupt,” said Petersson, of Andon Labs. “They just kept buying stuff. But now, it seems like it has more of an understanding of how much money it has.”
Lebedev, the first customer, said he appreciated the store's attempt to showcase AI’s capabilities but was glum the exercise was necessary. “I wish this experiment didn’t have to run,” Lebedev told NBC News. “I wish we lived in a world that was mature enough and had more democratic oversight over giant, very powerful AI companies.”
“This is not the technological progress I was promised,” Lebedev added. “I want technology that helps humans flourish, not technology that bosses them around in this dystopian economic hellscape.”
Andon Labs is quick to note that the store’s human employees are fairly compensated and given proper legal protection. “No one’s livelihood depends on an AI’s judgment alone. For now,” the company said in a statement. “We don’t pretend to have the answers here, but we want to start the conversation by publicly demonstrating that this future might be nearer than many think.”
For their part, Backlund and Petersson think the breakneck pace of AI development could continue or even accelerate in the coming months and years. If true, understanding how autonomous systems act in the real world — and how to make them safe for humans — will be critical.
“We want to be the provider of information, and then society and people who are smarter than us can decide what we want to do with this,” Petersson said. “We are not claiming that this is amazing for the world, and we’re not trying to optimize to make this a thing.”
Instead, the pair hopes the store provides a tangible AI experience for people who might not otherwise interact with the advanced systems. “The alternative, in our minds, is that it just happens before anyone has had time to think about the consequences,” Petersson said.
