The friction of diction
Speech recognition systems have long held out the potential to take the friction out of conversations, providing human-like customer interactions for businesses at a fraction of the cost. Unfortunately, reality has fallen short, at least in the world of customer support. Few things are as frustrating to a customer as repeatedly shouting “Representative please!” into the telephone, to an automated voice that keeps responding, “I’m sorry, I’m having trouble understanding you. Let’s try again.” Eventually, if you’re lucky, you’ll be connected to a live human, but it can take some doing.
I thought about this because Google has recently reported that its speech recognition technology is almost on par with humans. What this means I’m not sure, but coming on the heels of announcements in a similar vein from IBM and Microsoft, I imagine that, contrary to claims, there’s still quite a bit of room for improvements. Surely these will be made, but one wonders how much better the experience of talking to a customer service bot can get.
One reason for skepticism is that some business interactions between people just seem better suited to digital automation than others. The submission, processing and payment of an invoice—a fairly straightforward and routine process that’s all about accuracy—makes sense to streamline using digital processes that can take human missteps out of the equation. A conversation with a frustrated customer about an odd problem they’re having with a complicated new product is another sort of process altogether.
The reasons for this difference may have something to do with the inability of computers to understand a problem, rather than merely solve it. Computers, that is, don’t really care about problems, they merely solve them (hopefully), whether the problem is getting your information and connecting you to the right person, or telling you your bank account is overdrawn.
People, on the other hand, are capable of caring about your problem, and even if they don’t know how to help you, they can spend some time discussing it with you while offering a sympathetic ear.
That may be why invoicing and payments are so well-suited to digital transformation, and why customer support is more difficult. Because when you submit an invoice, you want it to be paid without mistakes, on schedule, according to set terms. Whereas when you call customer service with an aggravating issue, you’d prefer to get to someone who cares, someone who might even say “I’m sorry”—and mean it—which is something a computer just can’t do.
In other words, trust is earned through accuracy, speed and reliability in the world of payments, but may require something more in the world of customer support.
Ironically, as digital technology is enabling speech recognition, allowing computers to talk more, people are talking less, thanks to…digital technology. It’s well known that those who came of age in the era of cellphones are using them less for phone calls and more for texts, emails and social media posts. Partly in response and partly to cut costs, businesses have transitioned to, or ramped up, online digital customer support offerings. For frustrated customers, it’s one more way of easing the path to support.
But the biggest impact that digital transformation is having on customer support may be occurring outside of customer support, because of increased transparency. By way of example, one of the benefits of e-invoicing automation for AP departments is the time it saves them on unnecessary supplier phone inquiries. A supplier who can see the status of an invoice online no longer has to call up the buyer’s AP department to find out what’s going on with the invoice. The problem—“I don’t know what’s going on with my invoice”—no longer exists. And that has to be the best kind of support of all.