The tech geniuses have finally come up with something really useful for customer service representatives. It's about time. With the exception of some from a minuscule number of exemplary companies, I lost patience with those twerps a long time ago. But help is on the way. (Yeah, right.) According to an article that appeared in The Washington Post on October 18, 2006, some companies are beginning to use "speech analytics" and "emotion detection," technologies that were originally developed for security and intelligence eavesdropping, to improve customer service.

Speech analytics software evaluates and parses spoken conversations and transcribes them into written text. Rather than looking at just the literal words, speech analytics also allows systems to understand colloquialisms. For example, the article suggests that some of these programs would realize that "this is the last straw" is a negative phrase and not one uttered by overly insecure customers who feel the need to explain the rationale behind their orders for more tubes used to suck up drinks. Emotion detection adds to this capability by also recognizing the sentiment underlying the spoken words.

For the most part, companies employ these capabilities to analyze call patterns after the fact. They then use the results of their research to help improve their products and services. For example, the article suggests that telephone representatives don't always record why a customer is defecting to a competitor, even when the reason is mentioned in the conversation. Obviously, that's valuable information because it might indicate a problem that can be solved, possibly cheaply and easily. Don't ask me why—the answer, which involves weighty aspects of economics, finance, accounting, marketing, and organizational behavior, is far too technical and complex to get into here—but most companies seem to prefer things that they can do cheaply and easily, so it's particularly beneficial when the solution turns out to be such.

The new technologies can capture useful, unstructured information when the human operators don't. Then, with a complete database of computer-generated transcripts, including categorizations of emotions, at their fingertips, marketers can, for example, search for all conversations that include words such as "your," "product," "is," "a," "piece," "of," and "$#!%." (Use your imagination to replace the "$#!%" with your favorite expletive. I'm not permitted to publish those expletives here.) Looking more closely at the conversations returned in the search results might help the company to uncover a previously unknown defect in the product or in the way it is marketed and delivered.

Some companies are using emotion detection software to identify when a customer on the other end of a phone line is starting to get angry. For instance, the article tells of one company that uses its systems to take a baseline reading of a customer's emotion as indicated by his or her voice in the first 10 seconds of a call. If the software detects a divergence from that baseline during the rest of the call, it alerts a supervisor. Presumably, it is then the supervisor's job to find a way to perpetuate the myth that the company really does, in fact, give a damn about customer satisfaction.

That sounds nice, but don't get your hopes up for even the small dignity of having a supervisor jump on the line just as you are about to jump down the throat of the subordinate. Few companies are using the technology in this way because, as one person quoted in the article said, "There are not enough supervisors around to handle all these emotional calls." No kidding. That's probably not the only problem. The very few supervisors who are around at the more customer-unfriendly companies likely don't have sufficient emotional fortitude to survive the barrage of haranguing they would get from all of the red-hot, irate customers whom so many companies seem intent on cultivating these days. Unlike their subordinates, who get to speak to the occasional easy-going, dare I say even happy customer, the supervisors would have no time in their day to do anything but stave off the slings and arrows of outrageously disgruntled customers. The supervisors would be in straightjackets within their first few hours on the job.

If more companies start using this technology to accurately detect, in real time, the emotions behind my words, I expect that in the future I'll frequently hear the following words, or other words to the same effect, spoken in an exceptionally desperate, pleading voice: "No! Please, no! I beg of you, Mr. Klebanoff, sir, please don't kill me. I'm too young to die. I'll stop jerking you around and help you now. Honest!" Come to think of it, that doesn't sound like such a bad outcome.

Being a totally nonviolent person, I would never kill anyone or even slap them silly, which I believe the Geneva Conventions do specifically allow when you're provoked by a company's customer aggravation (aka service) department, but I must admit that I have often felt that way when talking—or, more to the point, trying to talk—to customer service people.

One of the problems I see with this technology is that companies are deploying it too late in the customer service relationship. If they really want to analyze my level of satisfaction or dissatisfaction, they need something that will measure the increasing intensity and abruptness of the force that I use to punch the keys on my telephone as I proceed through an infinite hierarchy of "press one for ..." menus.

How my telephone is able to stand up to the abuse I inflict on it after reaching the fifth menu level is beyond me. It's a testament to the ruggedness of today's phones. Just as an aside, writing the previous sentences led me to realize that I've never had to call a phone manufacturer to complain about its product. I wonder how that conversation would go. "Hello, Acme Phone Makers? Hello? Hello? HELLO? Is anyone there? IS ANYONE THERE? Damn phone!" But I digress.

Getting back to the matter at hand, if companies truly want to get a complete picture of my frustration, conversation analysis and telephone keypad punching measurements are still not enough. They also have to listen to my loud, expletive-filled bleating in the last few minutes on hold while waiting for a human to come on the line, just before I finally give up because I find it impossible to go a whole day without eating, drinking, or emptying my bowels and bladder. Then again, avoiding the former two does extend the time I can go without performing the latter. But I'm digressing and painting an ugly picture.

All of this new technology is wonderful, and I'm sure that enlightened companies will use it to significantly improve the customer experience. Of course, it will, no doubt, cost a lot to implement, and it will likely result in large disruptions in current work practices, but it will be worth it. Then again, as an alternative, companies could build products that work and are so easy to use that we never need to call. Oh yes, they could also think about picking up their damn phones in something under half an hour when I call for service. Hey, it's just a suggestion. It's not like I really expect them to do it or anything.

Joel Klebanoff is a consultant, a writer, and president of Klebanoff Associates, Inc., a Toronto, Canada-based marketing communications firm. He is also the author of BYTE-ing Satire, a compilation of a year's worth of his columns. Joel has 25 years experience working in IT, first as a programmer/analyst and then as a marketer. He holds a Bachelor of Science in computer science and an MBA, both from the University of Toronto. Contact Joel at This email address is being protected from spambots. You need JavaScript enabled to view it.. Emotion detection systems that compare a customer's changing voice patterns to determine when a customer is getting cranky wouldn't work with Joel. Cranky is his constant state.