After seeing this Wall Street Journal review, I recently acquired Emily Yellin's new book, Your Call Is (Not That) Important to Us.  As someone with a voice user interface design background, I'd be thrilled if anyone responsible for call center automation had a chance to read it.  I haven't gone cover to cover yet, but while I was skimming at the dinner table I read a quote that echoes a concept VUI designers are no strangers to.  It came from Frankie Littleford, the vice president of reservations at JetBlue. 

She said that "customers...are so Internet-savvy.  They really don't want to pick up the phone.  They just want self-service -- to go online and figure it out themselves.  More times than not, once a customer calls, they have already tried to make it on their own.  A phone call is reserved for those more intricate, unique, complex situations."  I know the sentiment isn't 100% true.  Many companies' deflection rates prove it.  For one thing, cell phones are so ubiquitous that sometimes convenience wins out over modality.   

I bank almost exclusively online, but If I'm out shopping and fall in love with a pair of shoes (again) then I can quickly get my account balance before whipping out the debit card.  I'm mandated to use our corporate travel Web site, but if I'm stuck on an airplane wondering if I'll miss my connection the first place I turn is my airline's flight status line for an estimated departure time.  Every contact channel has a time and a place.  Nevertheless, Frankie Littleford's generalization is a better assumption going in than the assumption that everyone is willing to use phone-based automation to accomplish their goals.

This got me thinking about how to identify complicated calls from the beginning.  Usually, menus are used for this segmentation.  Both simple and complex tasks are offered, and callers (ideally) choose the reason for their call.  After the selection, some callers are sent to additional automation while others are sent to agents.  Usually the routing decision rests in the hands of the company and those who designed the system -- and it is often influenced by the perceived difficulty of developing the automation.  Too hard to do?  Let agents handle it.  Easy to automate?  Send everyone to automation.  But that logic can fall short; just because something is easy to design and code and support, it doesn't mean it will meet everyone's needs.  

A few weeks ago, I was in the airport for a business trip.  The security line was no longer a single line, but was broken into three categories:  Expert Travelers (the longest line), Occasional Travelers and Novice Travelers (the shortest line). Trusting the logic of the Transportation Security Administration (please don't laugh) I hopped in the Expert Traveler line and watched with interest to see if people progressed at different rates.  They did!  In the end, despite the different line lengths, I would have reached an x-ray machine at the same time regardless of my choice.  Here was an example of people self-segmenting themselves based not on their task -- we were all removing our shoes and unpacking our toothpaste -- but based on how challenged we expected to be by it.

So what if a call center did that with customers?  Maybe instead of an account number, the first thing an IVR needs to ask is "Are you calling about something routine or is your issue more complicated?"  People aren't inclined to lie about a direct question, just like the Expert Traveler line didn't fill up with slow and confused passengers.  Callers who perceive their goals as easier can be rewarded with concise and streamlined menus.  Those who consider their issue complicated will find themselves more quickly in an agent's care.  This approach is not the same as offering an agent early on.  That's way too alluring for many people with simple tasks.  Instead, consider the strategy a way of asking callers to trust where you steer them regardless of why they're calling.  

 If the TSA can do it, why can't everyone?