Home » Blog » Tech Trends » Brewing Beer with Artificial Intelligence

Brewing Beer with Artificial Intelligence

posted in: FOOD Tech, Tech Trends


Care for an AI-le? A British Company is betting that machine learning technology can help brewers turn out a tastier tipple.


One thing that any self-respecting connoisseur knows is that beer is actually alive. Their flavour is constantly evolving, even after you bottle it, and changes again after it’s poured. This is one of the reasons why the micro-brewery movement has taken off in such a big way, as it gives fans of the hoppy stuff (and I count myself among those) a lot more choice.

Beer is actually alive. Their flavour is constantly evolving Click To Tweet

Golden AI being chilled to the correct temp

But is there a thing as too much choice? A small-scale brewer faced with limited resources and endless choices of what to brew, what ingredients to use, and what methods to employ might well think so. With so many variables, how do you negotiate that uncertainty and figure out what it is your customers want?

With so many variables, how do you negotiate uncertainty and figure out what your customers want? Click To Tweet

That’s a problem that has long  interested Dr. Rob McInerney, who wrote his Doctoral thesis at Oxford on making decisions under uncertainty. After finishing his PhD he then went on to start Intelligent Layer, a Machine Learning company. And it was in the East London WeWork where they were based that he first met 10x founder Hew Leith, a former M&C Saatchi director. In classic start-up style, they eventually realized – perhaps after a few beers – that they could combine their strengths to fill a gap in the market.

IntelligentX created the world’s first “self improving” range of beers Click To Tweet

Thus a mash-up of the two companies – IntelligentX – was created to make the world’s first “self improving” range of beers that constantly evolve in response to customer feedback data. The system uses machine learning (specifically reinforcement learning and bayesian non-parametrics) to improve itself.

Customers who try the beer can actually “talk” to ABI (short for Automated Brewing Intelligence) directly by using a Facebook Messenger bot, giving feedback on the flavours. The AI then crunches all that data and helps fine-tune the next recipe of the brew to best suit their tastes.

Customers who try the beer can actually “talk” to ABI - or Automated Brewing Intelligence Click To Tweet

If that all sounds a bit too much like having a terminator making your drinks (hey, I think that would be cool, but it’s not for everybody) never fear. ABI also has a bank of wildcard ingredients, like adding fruit to a recipe, which brings in an element of unpredictability and serendipity, all carefully factored and calculated in, of course.

Leith stresses that this isn’t about machines taking over jobs, but about working closely with brewers to supercharge their intuitive, artistic skills using AI.

“We believe the future is a place where A.I. augments humans’ skills. In this case we’re using A.I. to give our brewer superhuman skills, enabling them to test and receive feedback on our beer more quickly than ever before. This means we can respond to consumers’ changing tastes faster than traditional brewers.”

We’re using A.I. to give our brewer superhuman skills Click To Tweet

McIrnewey adds that in a situation where you don’t have millions of datapoints to train a deep learning algorithm (like Google’s Deepmind did with AlphaGo) it is better to employ a Bayesian non-parametric approach to use the information you have much more efficiently. This means effectively rewarding the system every time it makes a correct decision. “We also work very hard to learn as much as possible from the brewer directly, so this really becomes a joint effort between man and machine,” he concludes.

The IntelligentX team (left to right) Rob McInerney _ Siseley Coates-Harman _ Hew Leith _ Steve Silvius _ Lucy Britton

You also know you’ve got your hands on a seriously techie beer when they announce plans to open source all their recipes (on GitHub perhaps?) so that anybody can then recreate their favourite beers at home. “As each batch will be unique, there’s no need to keep the recipe under lock and key,” they explain.

You know you’ve got a techie beer when they open source their recipes Click To Tweet

The beers cost £4.50 a bottle from Ubrew and come in four main styles:

  1. Golden AI – From a classic British golden ale recipe featuring Styrian Golding hops
  2. Amber AI – derived from a British bitter, with darker appearance and stronger flavour, with a hint of grapefruit
  3. Pale AI – Hoppy beer derived from an American pale ale
  4. Black AI – From a classic porter recipe, it has a strong smokey flavour

The four IntelligentX beers available at launch

IntelligentX has quite ambitious plans for the future. Not only are they in talks with several Michelin starred restaurants about stocking the AI Beers, but they also have their sights on eventually winning a major brewing award such as CAMRA’s Champion Beer of Britain. So far the beers have already evolved eleven times, so it’s probably only a matter of how many generations it will take to hit that sweet spot of maximum optimisation.

AI will eventually be used to improve all sorts of physical products Click To Tweet

I can certainly see this being a big hit with the Tech City “Beer O’clock” crowd, but there is also a much broader consumer trend emerging here. McInerney and Leith believe this approach of using AI to improve physical products will eventually catch on with all sorts of things, from chocolate and coffee to perfume. Why be stuck with eating the chocolates you don’t like? In future we will all be able to get products fine-tuned to our personal tastes, and it won’t even cost the earth, as personalisation can be quite efficient, with the right tech behind it.

We’ll Drink to that, Cheers!


IntelligentX: The World’s First Beer Brewed by AI from IntelligentX on Vimeo.

Alice Bonasio is a VR Consultant and Tech Trends’ Editor in Chief. She also regularly writes for Fast Company, Ars Technica, Quartz, Wired and others. Connect with her on LinkedIn and follow @alicebonasio and @techtrends_tech on Twitter.