Skip to content

What the coffee pour over wars teach ecommerce about artificial intelligence



Join our mailing list

Signifyd regularly publishes free reports packed with business insights, commerce trends and data from our massive Commerce Network. We’ll only email when we have something meaningful to share, no more than once per week. And of course you can unsubscribe any time.

There is a lesson about artificial intelligence for all of us in the battle brewing among America’s baristas about whether any machine can make a cup of coffee that’s as good as one made by a human skilled at the “pour over.”

What do you mean this debate is news to you? No matter, it’s a perfect way to frame the tension between the roles of humans and machines in ecommerce and business in general. OK, if not the perfect frame, it’s a fun frame.

Some background: The “pour over” is a method in which a barista packs coffee grounds in a filter nestled above a cup and pours hot water over it (preferably with a flourish). It takes time — five minutes or so. Some snazzy coffee houses have replaced the pour over method with precision machines that brew up coffee some say is every bit as good. Instead of making nine coffees an hour the human way, the machines make 100 cups an hour, The Wall Street Journal reports.

You can guess the seeds of the argument. Pour-over proponents: Hand-made is always better. Baristas can adapt to each customer’s idiosyncrasies. Enjoying coffee is a human, communal experience. The machine backers: Humans are imprecise, leading to inconsistency and mistakes. Hand-made coffee is expensive. Customers don’t want to wait for their coffee. 

Coffee pour overs caffeinate human vs. machine debate

These stands are black and white — unless you take yours with cream. 

All of this would be trivial if it weren’t coffee we were talking about. Seriously, all of this would be trivial if the themes playing out in coffee shops weren’t also evident in other businesses as the world rushes to embrace or recoil at the notion of artificial intelligence and machine learning. In fact, it’s becoming clear that these black-and-white stands aren’t the best ways to think about automation.

It turns out artificial intelligence isn’t really about human vs. machine — or it shouldn’t be, anyway. It’s about human-plus-machine. About deploying machines that amplify human talent and finding humans who make machines better at what they do.

Joe Beninato is a big believer in artificial intelligence. He founded Banter, a company that develops chatbots for customer service. You’d think he’d be all machine. But no. He says the value of using chatbots to provide customer service is that the bots take on the many, less-complicated customer problems, leaving humans free to deal with the bigger issues.

“In our case, we can automate about 70 or 80 percent of the inquires that come in,” he says. “We don’t want to lose that human element completely. If you only handle 70 or 80 percent, that means that 20 to 30 percent have to get help somehow and so we automatically route that into a human who can then answer their question.”

There are plenty of challenges in ecommerce that machines can handle well most of the time — building personalized product grids, screening for fraud, sending delivery notifications, responding to support questions. But there are always special cases.There will be instances that the machine has never seen before.

Starbucks coffee sign in a window

A human merchandiser might want to override an algorithm in the interest of promoting a brand value. A particularly odd order online might include a set of unique circumstances that a machine would flag as fraud, but which a human could determine was perfectly understandable. A customer service chatbot might miss subtle — or not so subtle — cues in tone of voice or language that a human could pick up on and respond to.

The beauty of artificial intelligence is not that it takes the human out of the picture; it’s that it allows the humans who are in the picture to tackle higher-level problems. 

The answer is combining the strengths of humans and machines

The same is true in serving coffee. There are times when Arlette Keyes needs her caffeine and she wants it now, which is why she forgoes a pour over for the Starbucks drive-through on the way to work, the Journal story says. But, the story adds, there are times, like when Keyes has a day off, that she will hit Intelligentsia coffee for a pour over. That’s when she can enjoy the ritual and the barista’s conversation or possibly what Blue Bottle Coffee founder James Freeman told the Journal is an experience approaching meditation.

And yes, even with coffee, there is a human-plus-machine model. The Journal story describes a machine that can perform a pour over. The robo-pourer lets the barista pick the water temperature and the rate of the pour over, the story says. The machine is no faster than a human, says the Journal, but while it does its work, the barista is able to gather up pastries and take care of other customers.

And let’s face it, the robo-pourer may be on to something. For once you have your coffee, there are no higher-level needs in the morning than pastries. 

Pour over photo by iStock. Starbucks sign photo by Mike Cassidy.

Contact Mike Cassidy at [email protected]; follow him on Twitter at @mikecassidy.

Mike Cassidy

Mike Cassidy

Mike is the head of storytelling at Signifyd. A former journalist and a retail geek, he covers ecommerce and the way technology is transforming digital commerce. Contact him at [email protected].