Table of contents
The salami sandwich was a staff favorite at Partners Coffee, a multi-location coffee shop and roasting business based in New York City that ships nationwide.
“Every time we put it back on the menu, it never sells,” said Andrew Cosaris, Partners’ Vice President of digital. “We just imagine it to be a top-selling item, and so we keep trying.”
But the numbers don’t lie, and year after year, sales data defied the team’s imagination, despite its beloved insider status. Eventually, the numbers won and the sandwich left the menu.
“I feel really bad for the sandwich!” Cosaris told me, half-serious, half-joking.
This, Cosaris said, is one immediate promise of increased artificial intelligence put to use inside a growing quick-service restaurant business. It presents irrefutable facts with the proper history and context to help a human make a decision. Partners Coffee eighty-sixed the salami before it started using AI to help analyze sales, but the tech would’ve offered the same solution, balancing a human professional’s restaurant experience with the kind of useful data that makes decisions like these a lot easier.
What’s on the menu? Just the facts, please.
Of course, AI that’s built into restaurant tech systems isn’t the only way to access a particular restaurant’s or restaurant chain’s sales data. The team at Partners Coffee had long been able to access performance information; the data. But the way they’re accessing it now, through built-in conversational AI inside the point of sale — released to Square sellers broadly this fall — makes discovering, comparing, and acting on the numbers much simpler than it used to be.
Talking to an AI bot is the same as texting a friend or, in this case, the most knowledgeable coworker you can imagine, armed with historical data and other business details and able to process and analyze information fast in a way that’s easy to understand.
Whether or not to keep an item on the menu, for instance used to be a complex, emotional decision. Is the payoff worth the time and effort? Cosaris said, “I think every business hypes themselves up about a hero SKU, but if they aren’t comparing it to year-over-year, they aren’t necessarily knowing, oh, maybe that hero SKU isn’t actually the hero SKU that we thought it was … and we have another product out there that is doing better, and we should refocus our efforts.”
In effect, the system takes the emotion — those “gut feelings” — out of decision-making. For Cosaris, that’s a good thing.
“I don’t want my gut involved in things,” he said. “My gut is involved with too many other things. I want to make decisions based on the data we actually have. I want to remove all personal opinions from these sorts of queries.”
Promise and potential for AI on guest experience
AI can do much more than analyze sales data, though restaurants are still testing, and even dreaming about what it might be able to do. The stakes are high, especially as consumers seemingly pull back on spending at restaurants amid uncertain economic conditions.
A Deloitte survey released over the summer found 82% of restaurant executives surveyed expected to increase investment in AI over the next year. Most hope the tech will benefit the customer experience; 84% of operators surveyed for the report said they’re currently seeing high customer experience impact from AI tools. While, the survey didn’t delve into specifics, immediate heightened customer impact after debuting new tech tools is an ideal outcome for a restaurant investing in these tools.
It’s already happening. A 2024 report from Square found that, among restaurants that used AI inside their restaurants, nearly all were using AI chatbots or virtual assistants to help with guest relations. Half of AI users then said these tools enhanced the guest experience.
Restaurants challenge technology, particularly artificial intelligence, to prove it can make smart operational changes that translate to happy human customers. Its potential to transform hospitality, arming employees with relevant, topical, and important information in near real time transforms potential heavy lifting — or work that wouldn’t have happened in the first place — to realistic and easy-to-implement data.
An extra brain that’s happy to help
According to Cosaris, comparing year-over-year item sales — an exercise that required pulling and comparing several spreadsheets’ worth of information — was a particular challenge before their point of sale could respond in chat.
“I would have never run these reports,” he said. “I would’ve said, ‘this is too much work and I’m not doing it. It sounds like interesting information that would be helpful to someone, but I don’t have the capacity.’”
It’s a use case I’ve heard repeatedly; AI isn’t replacing workers, it’s helping them take on tasks once considered nice-to-have. It’s an additional coworker, one that just wants to help. In fast-paced service scenarios, like a coffee house, AI arms employees and managers with information they can act on immediately, affecting real change in near real time.
At Partners Coffee, the chatbot is happy to help, without complaint or bandwidth concerns.
“It feels exactly like messaging a member of my own team,” Cosaris said. He even treats it as such; throwing in such hospitable phrases as “please,” “thank you,” or even a, “hey, do you mind?”
The bots may be programmed to comply with your written prompts, but this is the hospitality business, after all.
