It’s no secret that IVR (interactive voice response) has never enjoyed great consumer support because design and deployment commonly fall way short of a decent customer experience. This has even led some senior decision makers to openly consider whether they should ditch their IVR in the cause of being more customer focussed.
While this is a notion I sympathise with, it still leaves me baffled. Surely the effect would cause even greater customer frustration? Unless of course staffing ratios were massively boosted to play the receptionist role that the IVR is usually criticised for doing so badly.
But somehow I don’t expect that to happen anytime soon. And let’s remember the underlying economics that drive the need for these call steering solutions. Automated applications cost about one-tenth of what it takes to hire a live professional communicator to do the same customer servicing function.
Seems to me there is a clear ROI here for brands that still use customer facing staff to verify customer identity. As a UK national average, this task consumes between 20-40 seconds per customer according to benchmark research from ContactBabel. Now scare yourself and do the arithmetic for a six hundred seat operation over a twelve month period.. You have to agree that’s a bonkers waste of money!
Speech recognition and its more secure version speaker recognition can offer much smarter service experiences than traditional IVR. Not just for the initial identification and verification workflow but for all sorts of useful self-service that cut out the wait for live help.
Both technologies underpin evolving solutions that are well worth keeping an eye on for their potential to cut cost and improve service. It’s worth pointing out to UK readers, who get don’t get out much, that uptake here has been at a snail’s pace relative to US and Australian adoption rates. Both territories boast successful and ambitious uses that work in real life. These are light years away from the archaic ideas still clung to over here about standards of voice self-service.
But we are where we are. So what’s to be done? The issue with IVR is real and has been a thorn in the customer experience for many years. By the way, if you have an interest in these technologies and how they are applied, then Dan Miller and the rest from Opus Research is an email or RSS subscription well worth setting up.
Social Media Or Better Analytics?
A radical solution to IVR aversion could be to go social media crazy. Pursue the vision of outsourcing customer services to customers themselves. Re-invent your online experience with communities and tribal knowledge services to facilitate peer-to-peer self-help. A few pioneering brands have been fast to get going on this and so, at some point in the future, should enjoy reduced inbound call centre queues earlier than others.
But for those brands whose social media response is just a tactical twitter patrol and an isolated Facebook page, the problem is this. While inbound call volumes remain as they are, call centres need some way of providing a scalable routing service as soon as customers dial that customer service number.
Another way to get in control of your IVR is better diagnostics. It’s both surprising and disturbing to find that many call centres don’t actually have detailed or up to date views of what is happening within their IVR workflow. The usual global abandonment figures only communicate there is a problem. Not what, where, why and how to fix it.
Happily new options are now appearing. For instance a US webinar being held this summer (2010) is themed around the topic of ‘Using Analytics To Measure Speech IVR Effectiveness & Potential’.
“…Measurements of IVR routing accuracy, self-service success, and dis-satisfiers encountered in the IVR provide a data-driven basis for evaluating new speech investments. And, for maximizing the value of existing speech systems…”
In other words, use speech analytics to track down the gremlins and start tweaking. Undoubtably this will raise the overall quality of IVR experiences over time. But to address the main question of this post, the future of voice self-service is going to be driven by factors other than brands getting better with IVR optimisation.
The Next 2-5 Years
The way we interact with the technology we use on a daily basis is about to change. Over the next five years, many consumer based markets are aiming to bring all flavours of automated speech to our attention and use.
For instance talking postcards are now part of your Facebook communication options courtesy of Voxcard. If you are one of half a billion Facebook users you can find out what that means and have a go.
All those remotes, cluttering up every room with a TV, are due to be transformed through speech interfaces. Voice control will become ubiquitous. In preparation, future models of PDAs, mobile phones and DVD/blu ray players are being expanded and upgraded. This will allow all functions to be accessed using voice replacing the need to use a keypad.
Think about your TV viewing habits. Remote control navigation grows harder as the choice of services and content continues to expand. L4 is one of the companies now pushing voice based control to coincide with the arrival of GoogleTV (due late 2010 in US) which will no doubt encourage our growing desire to consume everything as an online experience.
Alternatively, if you like the idea of being able to listen to your text messages rather than read them, TTS (text to speech) applications are to be bundled with that new phone you will be getting. Naturally you will be able to reply hands-free by dictating your text message. And the whole hands free thing will work just as well for anyone with a strong Twitter habit. Especially when are on the move or driving in the car.
If you think about it, voice and the mobile experience fit together well. We are naturally disposed to both talk into and listen to our mobiles/cell phones. Online searching through voice makes sense instead of keyboard navigation when it’s a more convenient option. Particularly for some of us big fingered users!
A number of phone apps, from ShoutOut to Dragon and Vlingo, already translate speech into text messages and e-mails. Bing and Google both have mobile applications that let you search the Web by talking. One in four searches on Android devices is done using voice search. A recent promo from Google shows off the capability nicely. [youtube]http://www.youtube.com/watch?v=gGbYVvU0Z5s&feature=player_embedded[/youtube]
Increasingly these services will be designed as multi modal experiences. For example the larger screen sizes of smartphones naturally lend themselves to both visual (text) and auditory (voice) prompts. But why bother to do that? Because it gets us closer to the full sensory experience of live interaction. As a result, we experience more natural ways of interacting with self-service routines.
One thing designers have learnt since the early days of voice self service is that live interaction is a far more subtle art to imitate than previously thought. Hence the aim of multi modal design is to try and replicate the flexibility of live ‘human to human’ interaction. For example this means that spoken, button-press and touch commands may all be freely intermixed as options to help you more easily control an application. In particular, direct text input, by button or touch, is always available as an alternative to speech.
All of this innovation is being nurtured in the slipstream of mobile’s rapid growth. In particular, the accelerating uptake of smart phones spreads mobile based voice activated services all the faster. This is because of their greater processing power to handle voice recognition relative to traditional handsets. And according some recent Ovum research, 75% of all UK contract phones are now smart phones!
And if none of that stuff tickles your fancy I’ll finish by introducing one of the true research heavyweights in this space. Professor Wolfgang Wahlster. One of the keynote speakers at the recent SpeechTEK Europe conference in London, he talked about the mega trends affecting this space. I’ll pick two that impact call centres.
The first is that solutions will start to emerge that can deliver age and gender recognition capabilities for personalized call center service. In other words age and gender based segmentation!
The second is a new level of speech recognition capability that can deal with overlapped, non-native, accented, and spontaneous speech and provide customer interactions as transformed text files in near real time. Another step closer to real time responsiveness!
Good Self Service Is All About Customer Centric Design
So, there is much pencilled in for voice self-service in the near term. However to make any real progress the lessons from traditional IVR need to be thoroughly learnt. One of the key design failings was how often users seemed to walk into a workflow dead-end prompting them to zero out in search of help. Hopefully multi modal designers recognise this as a core customer experience challenge and will turn the issue on its head by making the technology do the work and not the caller. There is plenty of great technology at hand.
If enough brands can do that so consumers enjoy a better than average chance of a positive customer experience using voice self-service, then the UK might finally enter the world of 21st century customer service automation.