Robots on the front line of Customer Service
“Obviously, no human being ever saw it.”
I recently saw a post from somebody who had received a canned response to a communication sent to a major internet company. The response was quite wrong: the person didn’t have a problem, but was simply making a suggestion. The response said something like, “I’m sorry you are experiencing a problem. Please contact our Customer Service at…”
I responded to explain what might have happened, but thought it was worth addressing here.
There is no reason to suppose that a human won’t see something you send in, just because it is a robot that responds to you. That might have been true years ago, but things are very different now.
Here’s a likely scenario:
First, you send in your e-mail.
Second, a robot reads it, and tries to figure out what you are asking. If they can identify it with good confidence, and can send you an answer, they will do so.
That’s part of what has changed. Robots are much better than they used to be both at understanding human language, and judging just how well they have understood it.
What happens if the robot can’t solve the problem?
One good response, and that’s probably what happened here, is to direct you to a human being.
This system reduces the cost of Customer Service. Suppose your question is, “When is such and such being released?” Do you really want the company from which you buy products or services paying a human being to answer that question…particularly a human being who might need to have specialized knowledge in other areas?
No, it makes more economic sense for a robot to answer that question…which in turn, can make those goods and service less expensive for you.
The robot in the example at the top of this post made a mistake…it mistook a suggestion for a request for help.
Think about it, though: isn’t it better for it to make that mistake, than to err in the other direction? Wouldn’t it be worse for it to think you were making a “no response required” comment, when you actually needed help?
Now, suppose that a hundred people a week are asking the same question: isn’t it important for somebody to know that, so at the least, the answer can be better communicated to consumers?
Absolutely…and I would expect contemporary Customer Service robots to do that. I’m sure they tally the number of times that the same question is asked, and red flag ones which seem excessive…and those “call-outs” are then reviewed by humans.
This “robot first, human second” strategy is going to happen a lot of places.
As I’ve written about before, it’s happening with surveillance systems in public places. AI (Artificial Intelligence) systems examine the security video, looking for things that seem out of the ordinary to them. When they detect them, they flag the video for human review.
This is a great way for robots and humans to work together.
Robots* (including computers) are much better than humans are at doing the same thing over and over again. Humans are not at all good at that…we get bored, our minds wander.
Answering the “release date question” is that sort of repetitive task.
Humans, though, are much better at making decisions. They are also better at understanding other humans.
What I’ve outlined lets us both play to our strengths…repetition for robots, decisions in unusual circumstances for humans.
Expect to see this everywhere. I can already take my blood pressure at the grocery store. It warns me to contact a human being, though, if something is out of the ordinary.
It’s also the way it works for many people at the bank. They use the ATM (Automated Teller Machine) first for simple, repetitive transactions. If something goes wrong, the machine may direct them inside to see a human teller.
So, don’t assume that just because a robot tried to help you, that was the end of it. They could be just part of the human-robot partnership in all of our futures…
* A robot is something created by humans (directly or indirectly) that performs tasks (autonomously or not) done by humans (or, more broadly, by other animals…a robot dog, for example, would perform work done by living dogs, including providing companionship).
The word may conjure up an image of a mechanical man, perhaps clunky and made of metal. The way we use the term at The Measured Circle, it would include software performing human tasks, and non-anthropomorphic devices like an answering machine or a calculator.
This post by Bufo Calvin originally appeared in the The Measured Circle.