What IVR Systems need: a little more human
CRM Buyer’s Erika Morphy hit the nail on the head with respect to call-in customer frustration today. And she did it with one line:
Like, why do I have to repeat my account information over and over again?
In her exploratory article, IVR Systems, Part 1: Are They Done Yet?, Morphy reviews the progress IVRs have made over the years, all while observing that IVR innovations tend to move at a relatively slow pace: “For one thing, innovations in speech recognition and IVR systems tend to creep along.”
Morphy proceeds to highlight some new tools intended to improve IVR systems, all while remarking that they are still trying to solve issues from 15 years ago–like accents, background noise and voice modulation. The innovations sounds great and pretty exciting, but I have to wonder, is that the way to go?
I believe that the reason that IVR designers are still dealing with the same quirks and issues as 15 years ago is for a very simple reason: we’re still just as human as we were 15 years ago.
Technology can only take us so far. I believe that:
- We can–and should–tweak and fine-tune our IVR systems over weeks, months and years
- IVRs should be seen as living, breathing, ever-changing entities, not as lay-it-and-leave-it affairs
- IVR designers and IT folks should meet monthly with call center operators to discover what the most common questions, complaints and issues are and to drive down to figure out what is causing any time of customer frustration, both with the product and with the IVR system itself
- The IVR should reflect the way customers want to use it, not the way technologists want to design it.
That being said, let’s go back to “we’re still just as human as we were 15 years ago.”
Technology can only take us so far. How do you program out a sneeze, an “um” or a Southern drawl without introducing less accuracy for other humans? I don’t mean to be discouraging, but I believe that the reason that the pace of innovation in speech recognition is seen as slow is because human beings can’t be programmed. We’re just too weird.
Has anyone else thought of making speech recognition more human, not more automated? Has anyone else thought of adding a larger human element to IVRs, not a smaller one?
And no, I’m not talking about adding more agents; I’m talking about having human beings check in where speech recognition fails. You and I can hear, “847-942… get down from there! down, boy!-6738” and know exactly what the phone number is. But there isn’t a speech recognition program today that could record that number correctly on the first try. So why not take advantage of our human ears and have people listen in to fix what speech recognition can’t get?
But I should stop here, because at this point I’m just pimping our product, and that’s not what the blog is for. Back to the main idea. Here we go…
I believe that it will be a long time before technology catches up with the complex patterns of human speech. In the meantime, instead of focusing on adding more layers of technology, I believe we should back up and consider adding a bit more humanity into our IVRs and our customer service.
What do you think?