TCS Daily

Talking to Robots

By Glenn Harlan Reynolds - December 1, 2004 12:00 AM

A while back, I welcomed our new robot employees. Today, I have more reason to do so.

I get my phone service from Bellsouth, and like lots of people we have multiple lines. Over the holiday weekend, the fax line quit working. I spent some time fiddling around, concluded that the problem was probably with the phone line itself and not with something left unplugged or otherwise misconfigured somewhere in my house, and called the Bellsouth repair number. As always in such circumstances, I steeled myself for a long wait, to be followed by an extensive phone conversation with a clueless (and, given the holiday weekend, probably sullen) "customer-service" representative, to be followed in turn by an even longer wait until the service technician showed up.

That's how it used to be. In fact, that's how it was the last time I had a problem like that. But we're living in the 21st Century now, and this time things were very different.

My call was answered immediately, by a friendly robot. Our entire conversation was handled by voice recognition, and after I had described the symptoms the robot said "I'm testing your line now." There was a brief pause, and it went on to tell me that there was indeed a problem with the line, and that a technician would be dispatched. I was off the phone in five minutes -- no interminable hold, no fractured call-center English, no delays.

Even more impressive, there was a technician at my house in 30 minutes -- on Thanksgiving weekend -- and the line (which had been damaged by the cable guys, natch) was fixed posthaste. Less than an hour after I made the call, the fax line was working again.

This experience reads like somebody's 1990 futuristic (and optimistic) scenario, but in fact it wasn't until later that it occurred to me just how big a change it represented from business as usual a few years back. And, following Mickey Kaus's first rule of punditry ("generalize from your own experience"), I want to point out some lessons that it might hold.

First: Better service. We often talk about replacing people with machines as an economy measure, and of course it is. The "robot" I talked to wasn't a shiny, clanking "danger Will Robinson" sort of robot, but simply a piece of software. It had development costs, of course, but the marginal cost of running it is close to nothing. As a customer, though, what struck me wasn't that Bellsouth was cutting corners -- it was how much better my experience was because of the robot. When service improves, it doesn't feel like corners are being cut even if someone is saving money.

Second: Job loss. Not long ago, somebody would have been at the other end of the line. He or she might have been unhappy to be there, fielding dull calls from unhappy customers on a holiday weekend, but there would have been a paycheck. Not any more. Now that person is doing something else, or is unemployed.

Third: Outsourcing. Jobs like this often get outsourced to places like India, but not this time. Instead, it was outsourced to a robot. In fact, the "job loss" above may well involve a job lost in Bangalore, or somewhere like it, not a job lost in the United States. (Judging by the U.S. current-account deficit, that's probably a good thing, though I imagine that Indians may feel otherwise).

Fourth: Efficiency. Although, as I mentioned above, a dull job was lost to the robot, I got to spend more time at my money-making job (writing a column for the Wall Street Journal, as it happens, though not on outsourcing, which would have been just too much of a coincidence) instead of waiting on hold. My wife, who uses the fax for her business, was able to send bills out to clients. And, overall, the sand-in-the-gears effect of dealing with customer service was reduced. Writ large, this kind of efficiency may well produce more new jobs in the aggregate than the automation replaces, though I don't know how you'd determine that. But it probably bears thinking about, because of the next point.

Fifth: We're going to see a lot more of this. The telephone companies are the cutting edge, but this technology is going to get more and more common -- and more and more realistic. The program I spoke with made no bones about being a robot, but I've noticed that these things are getting less obvious. At some point, it'll become very difficult to tell -- and some places won't want us to know. That raises a lot of interesting issues, but it's probably a topic for another column.

But it seems to me that we're crossing some sort of a divide in terms of the returns on investments in information technology, and that we're likely to see the benefits (and detriments) of these things appearing in a much more significant way. Just so long as nobody invents a column-writing robot any time soon...


TCS Daily Archives