How will artificial intelligence (AI) serve the coffee industry?

Image generated with AI-powered NightCafe Creator and human prompts

---

It takes lightspeed coordination between billions of neurons in the human brain for a barista to take your coffee order, prepare the beverage – and wish you a pleasant day.

But increasingly, there’s another sophisticated mind at work behind the counter. Artificial intelligence (AI) is no longer science fiction and is already being deployed by hospitality businesses around the world to optimise operations and deliver personalised customer service with greater accuracy than ever before.

Though the technology is in its infancy, AI looks set to revolutionise the global economy and even humanity itself. A recent report from digital consultancy Accenture estimated that AI could double annual national economic growth and boost labour productivity in developed countries by up to 40% by 2035. Meanwhile, a major PwC study estimates that AI could add $15.7trn to the global economy by 2030 – with $6.6trn generated through increased productivity alone.
 

“We use AI tooks across the business to increase our productivity”
Liam Farrow, VP of Digital, Bluestone Lane

For the hospitality industry, AI could bring a quantum leap in operational efficacy by enhancing productivity, efficiency, customer satisfaction – and ultimately profitability. Amid growing competition and costs, the potential to deliver seamless experiences and consistent quality at scale is simply too large to ignore.

What is AI?

The concept of ‘artificial intelligence’ has been around since the 1950s, when Alan Turing’s ‘imitation game’ tested the ability of computers to mimic human responses. The 1960s saw computing pioneer John McCarthy’s LISP programming language used to develop the first self-modifying computer programmes. Building on the advent of the microchip in the 1970s, today’s exascale supercomputers can process billions of computations per second to provide unprecedented ‘thinking’ power for AI programmes.

In short, AI can be defined as the simulation of human intelligence by computers and digitally connected devices that can learn, create and analyse with a high degree of autonomy. AI machine learning algorithms can generate insights from vast datasets, such as loyalty programmes and real-time transactions, that would be far too time consuming for manual analysis. There are also AI deep learning tools, which use neural networks for image recognition and time-series analysis and can optimise workflows and monitor footfall to forecast staffing levels. 

Starbucks' Deep Brew AI platform is enabling sophisticated personalisation on its in-house app | Photo credit: Starbucks

---

Natural language processing (NLP) algorithms are another application and can respond to text and speech prompts to display personalised product suggestions or guide customers through the ordering process.

The world’s largest coffee chain, Starbucks, has already developed its own AI platform, Deep Brew, which uses AI and machine learning to gather insights from Internet of Things (IoT) enabled equipment and app transactions to deliver personalised customer experiences, optimise store workflows and manage inventories. The platform can also aggregate revenue and footfall performance to determine the best locations for new outlets.

Meanwhile, Costa Coffee owner Coca-Cola recently announced a five-year strategic partnership with Microsoft worth $1.1bn to accelerate cloud and generative AI initiatives across its global business. Using Microsoft’s AI coding and language model, Azure OpenAI Service, the beverage giant hopes to ‘help employees improve customer experiences, streamline operations, foster innovation, gain a competitive advantage, boost efficiency and uncover new growth opportunities.’

A helping hand for hospitality?

AI has not yet reached the level of sophistication where it can make meaningful decisions in hospitality without human input. However, AI-enhanced automation can give operators another set of hands to reduce time spent on repetitive tasks, improve accuracy and increase productivity.

“Coffee is ready to be digitalised. If you look at other food industries, coffee is lagging behind,” says Andreas Idl, CEO and co-founder at Cropster. Founded in 2007, Cropster develops IoT software solutions for green coffee traders, roasters and operators to improve efficiency, connectivity and consistency.
 

“AI will permit better, data-driven decisions faster”
Scott Stouffer, Chief Sales Officer, Probat

For Idl, AI is the next step in automation for coffee businesses that want to extract deep insights from customer orders and equipment telemetry.

“The base tech has been around since the 1970s, but we now have large enough data sets and strong enough computers to run AI and train it. You can retrain AI every day with new data, so even if your equipment changes it will learn, adapt and become better over time,” he says.

Coffee roasting is one area where Cropster is applying AI to generate deep data insights – and even predict the future. Roasting Intelligence uses AI to predict when the crucial ‘first crack’ of the roasting process will occur, which “allows us to know how the temperature will likely develop over the next 120 seconds and show that in real-time to the user,” says Idl.

Peering two minutes into the future may seem like a small window of opportunity, but the AI tool can generate significant benefits for large-scale coffee roasters required to meet strict product specifications. This is especially important for large retail clients such as supermarkets, for whom even a minor inconsistency can lead to an entire batch being rejected.

“Roasters will know if they run into a potential defect or if they are on track for something they want. The accuracy is so high that we cannot improve it further without building a better sensor system,” Idl says.

Probat is another prominent coffee equipment business embracing AI. “AI will permit better, data-driven decisions faster,” says the Hamburg-based business’s Chief Sales Officer, Scott Stouffer.

Unveiled at the SCA Expo in Chicago in April 2024, Probat’s Roastpic is a smartphone app that uses AI to analyse green coffee quality based on images fed into the system. “The results are stored in the cloud and can be correlated to different roasts. This product has tremendous potential to change the way we observe and measure coffee through the value chain from farm to cup,” says Stouffer.

Graphic by John Osborne

---

The Human Touch

Despite AI’s undeniable business potential, there are also profound concerns about how the technology could impact jobs, livelihoods and privacy. A report from the Organization for Economic Cooperation and Development (OECD) estimated that 14% of jobs across 21 countries are at high risk of automation, including AI-trained robotics and admin tools, with food preparation workers and machine operators likely to be hit hardest.

Another study by McKinsey Global Institute has projected that up to 800 million jobs could be displaced by automation by 2030 and recommended that governments begin to retrain workers in at-risk occupations.

Meanwhile, huge corporations such as Walmart, Nestlé and Starbucks have come under the spotlight for their use of AI tools like Aware, an AI firm specialising in internal staff communication analysis that reports it has so far collected 20 billion individual interactions across more than three million employees.

However, for most coffee and hospitality businesses AI is intended to reduce repetitive and time-consuming tasks so that staff can focus on delivering customer service and quality.

“We use AI tools across the business to increase our productivity,” says Liam Farrow, VP of Digital at US boutique café group, Bluestone Lane, who credits AI tools such as Chat GPT for producing marketing copy and Gemini Advanced AI for speeding up finance modelling as valuable additions to the team.

“We recently used Chat GPT to validate the modelling for our loyalty programme. This delivered improvements in the value we could create in the programme in line with our goals. That task could have taken days, maybe even weeks, but was completed in hours by throwing all the different variables at the model,” says Farrow.

“I see the immediate upside in helping make the support resources more effective. Hospitality generally doesn’t allow for large support teams, so you need to be efficient. Removing tedious and mundane tasks or the need to call in external resources for things like copy writing means you can move faster and have more impact,” he adds.

Cropster’s Idl agrees that AI works best in a supporting role for coffee and hospitality businesses. He highlights that while AI can feedback data to a coffee roaster and even automate temperature and time control, manual labour involved at the roastery, such as loading beans into the roaster and setting the desired parameters will still need to be done by hand. For this reason, Idl sees AI as a tool for increasing quality and consistency while reducing waste rather than as a replacement for human staff.

“AI will not load your coffee. It will not blend the coffee – it will not even make the decision on how to blend the coffee – and you’ll still need to pull out QC samples,” he says.

That being said, AI is already weighing in on the sensory appreciation of coffee. In April 2024 the Brazilian Coffee Industry Association (ABIC) unveiled a new AI tool for use in cupping. Tasting evaluators input coffee-tasting notes into an application, which then uses an algorithm to generate beverage ‘styles’. The aim is to make groups of coffees with similar or appealing traits easier for consumers to identify, thus increasing the coffee’s marketable appeal.

“Our focus is on the consumer, so that we can communicate what they need to know at the time of purchase, to choose their favourite coffee among the different styles available,” says Camila Arcanjo, Master in Sensory Analysis and Quality Consultant at ABIC.

The user interface of Cropster’s AI-powered Roasting Intelligence software | Photo credit: Cropster

---

“I can see a future where we measure green coffee at origin and suggest how the coffee could be sample roasted to achieve the flavour profile the roast master wants. We will digitally measure the finished product and suggest roast and blend corrections to get exactly what the roast master desires. Finally, we’ll pre-programme the production machines to roast four tons an hour of “African Ice Coffee Blend” to be perfect in whatever application is chosen. We might even link to espresso machines to pull perfect shots,” says Probat’s Scott Stouffer.

That future is already closer to reality for Finland-based Kaffa Roastery, which in April 2024 unveiled its ‘AI-conic’ coffee blend. According to Kaffa founder Svante Hampf an AI model was given descriptions of the brand’s coffees and instructed to create a ‘new and exciting’ blend. The model “somewhat weirdly” chose to blend four origins – Brazil, Colombia, Ethiopia and Guatemala – instead of the usual two or three, however, Hampf said the project marks “the first step in seeing how AI could help us in the future.”

Getting to know you

AI can be a powerful tool to help boost staff productivity – but could autonomous machines ever take their place? Keith Tan, founder and CEO of Singapore-based Crown Digital is betting that the future of coffee shops lies in fully autonomous AI-enhanced service.

Tan began developing his fully-automated coffee concept, Ella, after encountering acute staff shortages at his first café venture. The AI-powered robot barista can brew 200 cups of coffee each hour using more than 300 different coffee recipes and currently operates at metro and airport locations in Singapore.
 

“With AI and automation mass personalisation is far more possible”
Keith Tan, CEO & founder, Crown Digital

“Coffee shops and cafés are very manual and require human labour, but it can take six weeks to train staff and they might only last eight weeks. At some point, you’re going to need to automate processes, whether through robotics or super-automatic machines, and AI is going to manage the data and bring unparalleled consumer insights at mass,” says Tan.

Crown Digital has used Eversys super-automatic machines from the outset and has built vast datasets for extraction times, grind size, input and output temperature across all locations. AI is now using that data to generate real-time insights to assist with demand-led stock management, time-sensitive order prompts, personalised suggestions, predictive maintenance and restocking.

Tan uses the analogy of being able to tap into the local barista who knows customer order preferences at a specific location and scaling that knowledge across cities, regions and even globally.

“Coffee is a very personal drink – it’s all about consistency and familiarity. That’s very difficult to achieve across multiple sites, but with AI and automation mass personalisation is far more possible,” he says.

ELLA, Crown Digital's AI-enhanced autonomous coffee concept | Photo credit: Crown Digital

---

How far should AI go?

AI-enhanced machines could one day deliver entirely autonomous hospitality experiences. The question, however, is whether this is desirable in an industry like hospitality where human connections are paramount.

For Bluestone Lane’s Liam Farrow, AI has clear limits. “It’s still very early for hospitality to look to AI and automation but operators should be looking at their business and seeing where there are opportunities to generate efficiencies. Smart upselling based on dynamic basket analysis to provide a more personalised experience is about as far as we will go right now. The human element of our business is the core of Bluestone Lane – so that must remain true,” he says.

Many business leaders say that knowledge is power, and in a world where the internet has made access to information ubiquitous, using AI to generate meaningful insights from vast data sets is the next step for computing and could be a giant leap for the business of coffee. Beneath the hype and headlines, artificial intelligence is already enabling coffee and hospitality businesses to achieve greater operational efficiency, precision and personalisation on a mass scale.

“I don’t know where AI, coffee roasting and Probat are going. But I do know that we will learn new things every month and new opportunities will present themselves. The result will be better data for quicker decisions,” says Probat’s Stouffer.

That logic should certainly ring true for the thousands of hospitality businesses around the world grappling with increased overheads, product costs and staff shortages.

For now, Artificial General Intelligence (AGI) – machines capable of learning independently from human input – remains firmly science fiction and will likely not come to fruition for several decades.

Nevertheless, artificial intelligence is moving at a rapid pace. At the time of writing this article, Chat GPT had only been widely available for around 18 months and already feels like an irrevocable part of daily life. Younger and digital native consumers are already demanding seamless hospitality and retail experiences that would not have been possible a decade ago – and AI can provide the solutions to deliver for the customers of the future.

With tech giants Microsoft, Apple and Google all pouring billions into AI research – and Elon Musk predicting that AI will be “smarter than any human by the end of 2025” – it is surely only a matter of time before the technology becomes as ubiquitous as an internet connection.

It may be nascent, but AI is already delivering tangible benefits for coffee and hospitality businesses, generating deep operational insights to enable quicker, more accurate and higher-value decisions. The potential for AI in hospitality is huge, so long as the industry keeps sight of its most important objective – forging deep human connections.
 

This article was first published in Issue 19 of 5THWAVE magazine.

Subscribe to 5THWAVE to receive each edition in print and digitally or sign up to our newsletter and be the first to read the latest articles and updates on World Coffee Portal research.