• 10.-12. October 2022
  • Exhibition Centre Berlin

21. June 2022, by Jan-Peter Wulf

Artificial Intelligence? – Head to the Backend of Bartending please!

Will robots really be mixing our drinks in future? Where does Artificial Intelligence make sense at the bar and where not? Our author and bar aficionado Jan-Peter Wulf gave some though to this matter. 

 © Shutterstock

The image video of the start-up MyAppCafé from Pfinztal in Baden-Württemberg is atmospheric by all means. Accompanied by some lounge music, a woman clicks on a cappuccino on her smartphone App, holds the resulting barcode under the scanner and the barista robot does its job: the gripping arm reaches for a cardboard cup, places it under the coffee machine, the liquid is poured and finally the desired “latte-art” lettering “Good Morning” is produced on the frothy top. But in the end, the woman is also standing all on her own in front of an empty box. “MyAppCafé” runs virtually without any staff. According to the image brochure the system only requires human support for an hour a day, for restocking and cleaning. Currently, there are already several locations in shopping centres – Hamburg, Heilbronn, Karlsruhe, Vienna. Is this what the future of (coffee) bars looks like?

Bar robots, really?

A barista robot of this kind could – strictly speaking – also be a bar robot and, needless to say, there are already some developments of this sort available. In a broadcast of the Pro7 science show “Galileo” from 2017 a metal colleague of this kind by the name of “Tipsy Robot” was introduced to the audience with the words “Met any barkeepers who know 500 drinks by heart but don’t chat you up?” Well guess what, dear Galileo: I have! And not just one. Many good barkeepers have 500 drinks in their repertoire. Not all of them are “pre-programmed” like that robot but many – in addition to a host of recipes that are stored on this veteran’s “natural hard drive” – are created à la minute, so to speak. And this repertoire is never-ending! How often have you sat at a bar – or a colleague’s bar if you, dear reader, are a bartender yourself – and a new drink was created simply by toying around with a new product sample. Maybe it was just a “one-hit” wonder, maybe it made to the menu. This is what you call creation – take that robot!

Artificial Intelligence 0 – natural empathy 1

And as regards talking up people – this is actually what going out to a bar is all about – putting the polemic aside for once. It is about the conversation, the exchange, the recommendation. The encounter! A bar is a third place in the meaning of sociologist Ray Oldenburg: a place of community that provides us with a compensation for and diversion from our first place ‘at home’ and second place the ‘workplace’– all the more today where places one and two converge in our home offices. What should I talk about with a robot? How is the robot to read from my eyes, my posture whether I am more inclined to have a ‘Bellini’ or rather a ‘Vieux Carré’? And is silence not more valuable if my human counterpart realises this customer actually wants to be left alone to enjoy their drink in silence? Artificial Intelligence 0 – natural empathy 1.

Wow effect wears off

Sure, there will be places where such futuristic cocktail robots operate. People will go to such places and order drinks. But they will most probably go to these places once and order one drink – to rarely or never return. Because that initial wow effect – of a cocktail prepared without any human interaction – actually wears off pretty quickly. It is not by coincidence that the said bar robot is deployed in the tourism Mecca Las Vegas and the barista robot from Baden-Württemberg in anonymous shopping centres – rather than in classic bars or sociable coffee houses.  

Digital cat faces?!

This wear-off effect also applies to the new service robots that roll through restaurants serving food and beverages. The idea of reducing waiters’ workload by delegating the carrying of heavy platters to autonomously driving artificial colleagues is not bad in itself. But the way these look these days is simply unsightly. A service robot whose screen face really resembles that of a “cute” cat? This is whimsical at best once but unpleasant in the long run. Let’s consider this again by the time a Steampunk version is available; or if somebody designs the vehicles like old-school bar trolleys.  

Analogue is sexy

I know, I sound a bit nostalgic but this doesn’t mean I think there’s no place for Artificial Intelligence at all in bars. Quite the opposite: it just has to be re-thought and integrated differently: less at the frontend, i.e. not in mixing and service, because these are the essential reasons for going to a bar in the first place. Just put yourself in the shoes of the customer: why should people working at screens surrounded by technology with few human encounters all day long also be confronted with this in their scarce leisure time? They want all things analogue! Pre-order your drink by App on the train for later at the bar – okay, maybe. But isn’t the anticipation, that watching the drink emerge, part of the delight? It certainly is for me and I don’t think I’m alone in this.


AI-based staff scheduling
So Artificial Intelligence should instead be used at the backend. Even today there are a great number of exciting tools and developments available for this. Take the example of staff scheduling: instead of assigning shifts on slips of paper and by hectic telephone calls or Whatsapp messages, when somebody drops out at short notice, digital tools that allow duty rosters to be generated sometimes weeks ahead are advisable for food service operators. Shifts are changed or re-assigned via the Staff App, without the boss having to call his crew one by one. What’s more: the (artificial) intelligence of these systems lies in the fact that they can produce forecasts for footfall, staff required as well as turnover on the basis of past sales but also using weather forecasts and other data. Over time the sales productivity constantly grows thanks to the learning system. “Gastromatic”, “Shiftjuggler”, “E2N”, “Papershift”, “Planday” or “Staffomatic” are just some of these solutions.

Camera peeks into the waste

Take another example like waste avoidance: how much fresh produce – citrus fruit, bar food leftovers and more – actually ends up in the waste bin? There are also smart tools to control this: cameras are installed above the waste bin and scan what is dumped while the waste is weighed at the same time. This allows the waste to be sorted in a “clean-grade” way. Such systems called “Kitro” or “Winnow Solutions” are primarily geared to food service kitchens but they could also come in handy and save money at bars in the long run (especially if their prices comes down).


Pinching drinks menus 2.0

A further example is trends: where do I find new ideas for cocktails? Of course, you can do your own scouting to a certain extent; paying visits to other bars (in other cities) is always a good idea. Cross your heart: who hasn’t taken a photo of a drinks menu there or just gone ahead and pinched it? The same aim is achieved – albeit in another dimension – by the “Tastewise” software from Israel: it reads (drinks) menus. Hundreds of thousands of them. And it analyses millions of food and beverage posts on Instagram and the like – including the likes and rankings. This is nothing but “bar big data” – a giant data repository from which users can retrieve their own results. Which drinks are “hip” in San Francisco right now? In London? In Lausanne? Are there any ingredients that are used particularly often in a specific region? Where do I find cool creations with Kombucha or water kefir? In their book “Food Code”, which is well worth a read, authors Olaf Deininger and Hendrik Haase conclude: “Tastewise” could develop into a valuable partner for restaurateurs because it identifies a “demand not yet discovered”.

Strength lies concealed

AI at the bar scores points “in the shadows”, for instance in areas like sourcing and ordering, calculation and pricing, supplies and pick-up, reservations (with capacity-utilisation and sales forecasts rather than just a digital booking list), in areas like searches (Google) and Social Media, staff management and digital recruiting, payment and book-keeping, stock-taking and the oh so “popular” HACCP documentation. For all of these processes beyond shaking and serving there are solutions already available today to help firms better control their processes, operate at lower costs and free up some time to be more creative and communicative – because they outsource boring or annoying jobs to technology. Once again: why give away the best bit about barkeeping, the creative and communicative element of all things, to robots? This would be a lose-lose situation, for the team and for their customers. AI becomes a win-win solutions once it frees up time for the essentials.  

Bars need interoperability

What is still lacking today is proper connectivity between the different systems, including central control via a dashboard. Having a separate App for each area, a closed system, is clunky and ineffective – interfaces for one consistent process are better than silo solutions.  AI needs API (Application Programming Interface) – and there is still lots of room for improvement. The buzzword reads: interoperability: smart systems that interact with each other. This is when Artificial Intelligence will become really exciting for places like bars.