Automation is supposed to help, not hinder

It’s all Henry Ford’s fault. While it’s almost certainly true that if he hadn’t innovated the production line and interchangeable parts, someone else would have, he stands squarely in the gun sights of history when we rail against technology making humans irrelevant.

He saw that robots and automation could produce a more uniform product more efficiently, and we’ve been off to the races ever since. Computers only make it worse. Thanks to Bill Gates, even before the epidemic of big data, computers and the internet have been tried and convicted of killing the middle class, newspapers and, counter-intuitively, porn, via a variety of methods.

But the first one, the middle class, is the one I want to focus on. It’s beyond true at this point that people have lost white-collar jobs to computers. As any 10 minutes of MAD MEN will tell you, there used to be entire departments engaged in activities that today are done by one person or, at most, one team. Things like secretarial pools (for typing), mockup artists and even broad swaths of accounting have been felled by three words: Word, Photoshop and Excel.

But for the most part, that’s actually OK. Computers are designed to and should be used for streamlining everyday tasks, allowing people to work more efficiently and (because all things must have a Legitimate Business Purpose) even saving the company money by consolidating the number of employees to produce a given widget.

These are what we’ll call sensible (though regrettable) redundancies. But the problem with technological innovation is that we think that any problem, with enough sufficient amounts of tech wizardry thrown at it, will disappear.

The flaw with this philosophy is that, much as with medicine and side effects, sometimes the troubles with the cure are worse than the problem it was trying to solve.

It’s 5 p.m. You’ve come home after a long day of work and, according to Amazon’s website, your brandy new Shiny should be at the door. Amazon queried the UPS database, which confirmed that the driver had scanned the barcode on your package as having been dropped off at your home.

Yet, despite looking on the porch, peeking behind the rosebushes and checking with your neighbor, it’s nowhere to be found. Time for the phone tree.

Everyone’s dealt with phone trees. They do make sense, to a point. Why on earth would you route every single call through one (or more, depending on the size of your organization) person, who would then have to manually shift them off to the appropriate extension?

An automated greeting with options to go through for finding the person you want to reach makes perfect sense in a number of scenarios. But right now you’re waiting for a package that says it’s been delivered, even though it’s clearly not. And when you call up UPS, you damn well know that you don’t 1) Want to Ship A Package, 2) Track A Package, 3) Schedule A Pickup, 4) Inquire About Freight Services, or any of the other options the robot gives you.

You could try Tracking The Package. But you’ve already interacted with UPS system. We know that the system is wrong; it thinks the package has been delivered, when the package hasn’t been delivered. The problem is that the system has no conception that it could be wrong. All it’s ever going to be able to tell you is that the package has been delivered.

Naturally, you started mashing 0 the minute the robot asked whether you wanted to converse with it in Spanish (automatonic show-off). 0 is frequently the magic number that tells the system, “Sorry, I need to talk to an actual human being because you’re so arrogant you can’t even admit the possibility that you could be wrong.”

(Of course, the very first thing the helpful representative does is query the computer so she can tell you, “Sir, this says the package was delivered,” adding a third layer of the same information confirming itself, but that’s also another rant for another time.)

This is a design flaw, a self-reinforcing feedback. The system tells the website you're wrong, so when you call to inquire it checks ... the same system, and it of course agrees with itself. And the reason this is a problem is because the implementation of this automation actually makes the jobs (and lives) of humans harder. We’ve so completely bought into the superiority of computers that, faced in a real-life situation, we almost always take their word over that of a human being.

Consider how many times someone’s complained about the technology where you work. Is the software you use every day to do your job completely bug-free? Is it even designed to do the things you’re forced to do with it? Know anyone in the food service industry? Ask them about their Point of Sales system.

Think about all the customer service interactions you were involved in from the buying side that included faulty technology. How often has the employee said, “Oh, that’s clearly wrong, let me fix that.”? The best-case scenario in that situation is that someone gets sent to go check that you were not, in fact, lying when you said the shirt was on sale even though the computer didn’t realize it. Or everyone gets to cool their jets while the manager wanders over from the back of the store to enter the special “override” code that forces the computer to accept the input of the human being operating it.

All it boils down to, essentially, is that these companies trust their computers more than their employees. (Which points, frankly, to absolutely terrible HR work.) This makes sense if all you care about is hiring people you can pay a pittance who will do the bare minimum, and rely on the computers to police everything. It falls apart somewhat if you actually care about your customers not hating the experience of going to your store.

To a certain extent, they’re extending Ford’s maxim: Using computers gives them a more reliable outcome. The problem is that they don’t bother to alter course even when that outcome is awful, because they believe that hiring the people to do the task properly would be difficult, expensive or not worth the money.

Thus, homogeneity is prized over efficacy. And that’s their prerogative, I guess. After all, everything must have its Legitimate Business Purpose, and there’s no Business Purpose more Legitimate than “it costs me less money today/this month/this quarter.” And perhaps the giants of various industries (Amazon, UPS, Walmart etc.) are so entrenched — or they’ve devolved everything to the commodity level so that price is the only differentiator — that they’ll never have to worry about the upstart who innovates on service and providing a user experience that’s actually pleasant for the user. Just ask Microsoft.