Restaurants lean on AI to cut waste and reduce costs

The hospitality sector is known for emphasising the human touch. But behind the scenes at busy restaurants, artificial intelligence is playing an ever larger role in bringing freshly cooked food to guests’ plates — and easing the growing strains the industry faces.

Specialist AI tools for assessing pantry inventories, tracking expiry dates, taking orders, checking recipes and managing reservations are “giving back time to restaurant teams,” says Julia Gilbert, managing director of the Fallow Group, which owns London restaurants Fallow, Fowl, and Roe.

For Gilbert, the tools could reduce staffing costs and minimise food waste — as well as freeing up the time needed for waiters to tempt diners to go for “that extra drink or extra coffee.”

Finding new ways to save both time and money is crucial for restaurateurs facing rising costs, as well as intensifying competition from takeaway operators and supermarkets.

In the UK, hospitality businesses have warned they could be forced to cut staff and reduce trading hours, as rising energy prices coincide with increases to the minimum wage and business rates that came into force at the beginning of April.

Still, the sector has so far been one of the slowest to adopt AI — in part due to reliance on outdated technology platforms, a shortage of staff with AI skills, and a desire to protect human connection and service, according to a report last year by PwC.

Nearly one in three hospitality businesses in the UK did not yet use AI tools as of March, compared with just one in five across all small and medium-sized businesses, according to a survey by ChatGPT-maker OpenAI.

Fallow’s restaurants are “at the very beginning stages” of trialling different AI tools on handheld point-of-sale platforms, with features intended to minimise the steps required for waiters to take orders and answer diners’ queries. The aim is to free up time for staff: “They can converse with the table while they’re pushing the order through [and] it actually allows for more time for the waiter to make eye contact,” says Gilbert.

From the beginning of May, the company plans to roll out All Gravy, an AI-powered platform designed to support internal team communication.

The company’s recipes will be loaded onto the platform, so staff can use quick natural-language queries to check for allergens and recommend suitable dishes. All Gravy’s layout is similar to that of Instagram and TikTok, meaning that it is “very easy for the particular generation of servers that we have to navigate that system,” adds Gilbert.

Another tool uses AI to scan photographs of fridges, record inventory and calculate the volumes of fresh supplies needed to match diners’ demands, helping to cut down waste by ensuring the restaurant buys only what it needs. Gilbert initially doubted whether Peckish, the system she’s trialling at Fallow, would cope with the restaurants’ “very extensive stocklist” but says that so far it is “doing very, very well”.

The Evolv Collection, the restaurant group behind London institutions such as Bluebird and Coq d’Argent, is also developing an in-house stocktaking and ordering system to help chefs spend more time cooking, rather than on administrative tasks.

Evolv chief executive Martin Williams says that “historically a chef would be in [the] fridges writing down . . . stock,” before typing it up on a computer. Using the new technology, a chef can just speak the orders into a phone and the system will automatically place them.

Williams says Evolv is also looking at using AI to assist with tracking and forecasting trends in consumer preferences — including by monitoring social media and search trends — as well as planning wages and suggesting new restaurant designs. But he warns that hospitality businesses should be selective about how they use the technology.

His restaurants have so far avoided using AI to generate advertising material, for example, “because you lose that tone of voice with it,” he says.

Anything “obviously AI-generated” is likely to put off prospective restaurant guests who prize authenticity and trust — especially when they’re planning to spend heavily for a special occasion — adds Williams.

An AI-powered reservation centre does, however, speak directly with customers once they’ve decided to make a booking. “You won’t know that you’re talking to an AI-generated telephonist,” says Williams, “but you are”.

The aim [of implementing AI] is to streamline our in-house processes so the teams can just focus on serving

Even a diner’s restaurant selection may increasingly be shaped by AI. Online restaurant reservation service OpenTable has added AI voice systems for customers, including PolyAI, and last year launched an AI-powered “Concierge”. Now embedded across more than 65,000 restaurant profiles around the globe, the tool scrapes menus, reviews and descriptions to offer speedy answers to queries from prospective diners.

The company also uses AI to help restaurants respond to reviews, saving them time and money, as an industry with “tight profit margins and limited capacity” increasingly views AI as “essential infrastructure rather than a nice-to-have”, according to OpenTable’s chief technology officer Sagar Mehta.

Ultimately, however, Fallow’s Gilbert says that hospitality businesses can only succeed if they use AI to emphasise, not suppress, their human employees.

“We hire for big personalities and we want them to just be interacting with the guests and not stressing because the till has gone down or they need to scribble down notes,” she says. “For us, the aim [of implementing AI] is to streamline our in-house processes so the teams can just focus on serving.”