TuVox Speech Central

I’ve written at great length about how you can have the “best of both worlds” — cost-saving automation AND high caller satisfaction.

I’ll concede that there are millions of ways to implement automation poorly and there are relatively few ways to do it right.  It’s a challenge, to be sure.

So I was happy to read an article at Next Generation Power and Energy that talked about exactly that.

“We’ve seen customer preferences shift in accepting technology over the years. Customer feedback now clearly shows that in many cases, customers prefer using technology for some transactions when it’s designed well.” says Tucker Mann, Vice President Customer and Market Services, Progress Energy.

It’s refreshing to see a public utility out in front of this issue.  Utilities have gotten a bad rap for their customer service, and now it seems companies like Progress Energy are leading the charge.

Kudos, Progress Energy!

There’s little doubt that personalization (automatically delivering relevant information) and customization (controlling the display) have been a boon for surfing the web. We get more relevant information, find things more quickly, and are more likely to stumble (or Stumble) upon things of interest.

This works for the web experience because we get the thing we’re looking for immediately, and the elements brought to us via personalization are unobtrusive. For example, on the Google search results page, the main results are in the center of the page, with relevant ads at the top and on the side. On an Amazon.com shopping page, the item we chose is at the top of the page, and recommendations/substitutes are below.

Note that in both cases, the design allows direct access to complete the desired task – getting information or completing a transaction – we’re interested in, and puts the other information nearby – in case we need it.

What about the telephone?

The key difference from a communications point of view between visual and audible content is fast random access vs. linear, fixed speed listening (delivery). Audio requires attention, where graphical content is less demanding. It’s ok if your eyes wander a bit while you think and return to a page, but if you get distracted while listening you often need to listen again. This attention makes audio a bit more involving and more personal than a pure graphical presentation. Both these qualities – linear delivery and the involving nature play key roles in the use of personalization.

Arguably, the most critical factor to consider when using personalization techniques on the phone is the relevance of the information you provide. The relevance test is more important for callers than for web visitors due to the qualities of each channel. Highly relevant personalization strategies will add value by shortening calls, increasing automation, and providing increased customer satisfaction. However, if the relevance isn’t high enough, you’ll be presenting information to the caller which is potentially “in the way” of the caller’s objective. If you have any information to present which may not be relevant, be sure to consider presenting it only after the caller has completed their primary task.

Be sure to check out the new TuVox personalization module, a great way to get together the data and rules you need for a highly relevant set of offers.

Driving and any other activity shouldn’t go together because it’s really not multitasking but “distracted driving.”

Yet in our busy lives today, we are all trying to squeeze out that extra few minutes while we are driving. For example, while driving to work,

I want to make my credit card payment since it is already overdue by a day

or the stock market is crashing and I need to enter my order to cover my 1,450 cubes (QQQQ) short position?

or I am trying to find the nearest office products store to buy the special lamp we need for our LCD projector

Frequently, when I reach my credit card provider or my stock broker or my favorite office products store on the phone, I am served by the so-called all-efficient DTMF-IVR. “Thanks for calling (Your) Bank credit card. To ensure quality service and accuracy, your call will be recorded. Please enter your 21-digit account number followed by the # sign now…..” Now, if talking on a mobile phone without a hands-free device is considered distracted driving, I think keying in information using “touch-tone” while driving would rate as extremely dangerous! And considering that mobile phone keyboards continue to shrink in size – its almost impossible to key in numbers accurately while driving!

In California, beginning July 1, we all have to use hands-free devices – no more Touch-Tone while driving! Click here for more information.

I think businesses that serve me and you – our credit card bank, our stock broker and our office products store have a golden opportunity to help us here. Adding natural language speech automation to the contact center can serve us anytime (24 x 7), anywhere; provide a pleasant caller experience; and enable the service in a manner where we don’t need to key in information, we can just speak it.

That’s why I say “Don’t Touch-Tone and Drive, Just Speak It!”

The recent issue of Business Week (3/3/2008) has an update on the “Get Human” initiative started by Paul English in 2005.

For anyone living under a rock during 2005-2006 when Paul was making the rounds of NPR and MSNBC, here’s what GetHuman is — you go to his website (www.GetHuman.com) and it lists a few hundred companies’ customer service numbers and tells you what to press or say to speak to a living, breathing person on the other end of the phone. Eureka!

After the initial novelty wore off, Get Human morphed into a proposed “standard”, with the half-hearted and opportunistic support of Nuance and Microsoft.

So here’s my take on this thing. When I read the Business Week article, I got the sense that the author (Jena McGregor) started out with a premise — Get Human is dead! – and pretty much wrote a page to support that preconceived notion.

If you actually go to GetHuman.com, you’ll see that a transformation has taken place since 2006.

Although the idea of a standard never took hold, the site now has a new focus. In addition to the IVR cheat sheet, there is a new “Consumer Rating” column, kind of like a Web 2.0 vigilante version of the Better Business Bureau.

And (hilariously), there is a column of affiliate web ads on the far right, implying that Get Human is paying its web hosting bills by driving traffic to the very companies they are exposing. Genius!

Angel.com’s Blog has an interesting take on the momentum loss of the Get Human movement. Ahmed Bouzid makes some very valid points about how the IVR/ACD vendors need to carry the flag to re-energize the campaign.

While that’s probably a sound approach, I prefer a market-driven strategy. In certain situations, people will always want to talk to a real person. Companies understand that. In other cases, the IVR is so unbelievably bad, people bail out because of sheer frustration.

Eventually, killer automation apps like the American Airlines IVR will prevail, and the lumbering old touch-tone dinosaurs will eventually die out.

And on that day, we might not need to “get human” any more.

TuVox Podcast:

Since last summer when we launched TuVox Speech Central, people have been sending me stories about really bad IVR and ACD systems that they call into.

I’ve been keeping a running list of some of the most offensive (and untrue) statements ever uttered by these abominations of automation. Here are a few worth noting:

1. Your call is important to us. Really? If it was truly important to you, I think you’d staff adequately, or at least give me an automated system to use.
2. For quality purposes, your call may be monitored or recorded. I’ve been hearing this little gem for years, and I’m still waiting for the quality to show up. A better way to word this might be “We record. Don’t sue us.”
3. Please listen closely, as your options have changed. No, they haven’t, and please take this annoying little message off its permanent place at the front of your menu. It’s delaying me from actually listening closely to the options, which I repeat, are exactly the same as they were the last time I called.
4. Did you know we have a website? No kidding. Welcome to 1995. Let me guess: its www, then your company name, then dot-com. Note to ACD administrator: please delete this non-informative announcement as soon as you read this.
5. Your call will be handled by the next available representative.Absolutely not true. The next available representative will handle the guy that has been waiting longest. Think about it.
6. Due to unexpected call volume… You can use this excuse probably once a year. Why, then, do I hear this message every time I call your call center?

Well, that’s all I can think of for now. Please let me know if I forgot any. And if you are responsible for an ACD or IVR, I would suggest killing some or all of these useless clichés of modern life.

I will pay you cash money.

TuVox Podcast:

Recently, I came across a video making fun of voice automated apps in an unusual service automation scenario: confessing your sins.

The video gently nudges the issue about how callers are at the mercy of rigid and badly designed voice apps.

Sadly, in real life there are lots of badly designed voice apps where caller are faced with no choice but to use the apps no matter what.

My personal experience with a few commercial apps is that I felt I was trapped in the voice system trying to do things the system’s way. All of us know that a bad voice app could be detrimental to caller adoption, potentially affecting customer loyalty.

I am not going to dwell on VUI best practices. A lot has been said about VUI best practices and is essential to all voice apps. But companies that spend thousands of dollars on designing cutting edge system fail to take into account customer feedback –which is the most critical part to the success of the app. Companies miss out on simple things like Wizard of Oz testing and caller surveys which help gauge caller perceptions and system adoption.

There are a few independent companies such as (www.vocalabs.com and Sterling Audits) that help you methodically audit and benchmark usability of your customer service. Those of you who want to follow the scientific benchmarking approach I would recommend following two books: Gourmet Customer Service and “How to Audit and Benchmark Your Voice Response System”.

So in closing, referring to a quote in the video, remember that if your users are complaining about your automated applications they are probably already in hell.