In the age of Siri, what are the new best practices for speech recongition in the modern call center?
It seems sometimes as if call centers haven't changed in the last 20 years, doesn't it? There have been innovations, to be sure: computer-telephony integration, integrated online chat and cloud-based architecture, to name a few. But has the experience of calling a call center changed all that much?
For most callers, contacting an enterprise for support or service still feels basically the same as it did 20 years ago. The customer experience doesn't seem to have improved substantially over the last few decades, despite a plethora of advances in the consumer space that have heightened connectivity, community and sharing.
In particular, automated speech recognition is a phenomenon that consumers love to hate. It's one of the tools that is rarely noticed when it functions smoothly but is always derided or even cursed when it doesn't. And some of the innovations of the last five years have added a twist: iPhone's voice command personal assistant Siri has gained acclaim in speech-recognition-related fields, leading to heightened customer expectations in the call center.
Let's take a look at the capabilities of Siri and how they affect best practices for the modern call center. Apple's virtual personal assistant has been an integral part of iOS since iOS 5 and was introduced as a feature of the iPhone 4S in October 14, 2011. While billed as having a natural language interface, Siri's primary strength is in its ability to follow commands related to smartphone application functionalities. Sending an email, setting an alarm, calling your spouse and getting directions are all squarely within Siri's wheelhouse. Most other queries will be directed to a web search for information. The beauty of Siri lies within the walled garden of its limitations: users know what their smartphone applications can do, so most voice commands are easily recognized by Siri.
The promise of Star Trek: an elegant voice channel
So how do these two tools affect the modern contact center? It's all about the promise of Star Trek. The dream of Star Trek was the elegance of saying "Computer, shields up!" and having the command carried out instantly. And that dream of an elegantly effective voice channel is still alive today: in a Forrester survey, 79% of respondents reported they still prefer the phone to other customer service channels. Ten percent actually reported preferring using the automated speech recognition system. That's 89% that choose the phone over social media, web chat, email and other channels. And the same study found that 84% of customer service interactions still take place over the phone, through live agent interaction or automated self-service. (The distant second category was email at a paltry 10%.)
While the field of customer service channels has expanded to include web, email, social media and SMS, the popularity and promise of speech-recognition-based tools such as Siri remind us that customers are still pursuing the dream of Star Trek: a voice channel that works with natural speech. While no speech recognition tool is there, yet, the promise of natural language processing is a tool that will easily and without question be able to process complex commands. The goal is to replace the stilted and linear voice processes with an elegant solution that understands and implements complex commands, such as, "Upgrade my flight to Japan tomorrow using my miles. If it's under $500, just do it."
The rise and refinement of the voice channel
The real value of modern tools such as Siri lies not in their ability to access increasingly large databases but rather in the refinement and promotion of the voice interface itself. Despite all beliefs to the contrary, customers prefer the voice channel. They just want it to work.
However, the field of speech recognition is shrinking rather than expanding. Few players remain: IBM, Google Voice, Nuance and Microsoft are the only companies still working on speech recognition and natural language processing at all.
Another wrinkle: the 80% shift
Add to that phenomenon a shift in the nature of inquiries to the call center due to the plethora of additional channels. For many years, common wisdom in the contact center followed the Pareto principle: 80% of callers are contacting the enterprise with the same, easily-answered, basic questions. Only 20% of callers had really tricky, complex issues that fall outside the frequently asked question (FAQ) paradigm and require human assistance, states the model.
But thanks to the expanded channels now available, that is no longer the case. Callers typically attempt to solve their issues online first and use the call center only when the solution isn't readily available within the frequently asked questions. In a recent study by Corporate Executive Board, researchers found that a full 57% of inbound calls to a service center where from people who had tried to self-serve first. That means that those 80% of FAQs are being answered through self-service on other channels. Therefore, it's the complex questions—formerly the 20%--that now make up the bulk of call center interactions.
How to modernize your call center
So given these three phenomena that the majority of customer inquiries are through the voice channel, that Siri has increased expectations for the voice channel, and that additional channels have shifted the content load to primarily complex questions through the channel, how do you modernize your contact center?
Don’t assume that the old 80/20 rule still applies. Most of your callers will have attempted to self-serve online first, so it's more likely that their issue falls outside the FAQs.
Find out what your top 100 queries are. Don't assume you know why your callers contact you; rely on actual data. And don't limit to the top 10 queries anymore—that's so 1990. Instead, compile a list of the top 100 queries your front line agents address daily.
The Long Tail power law
Train for the long tail. Since the queries are becoming more complex, you'll discover that among your top 100 topics, there is significantly more variety than there was ten or even just five years ago. This means that instead of training your frontline agents to answer three basic and five additional queries, you'll need to train them to answer all 100 queries.
It's not either/or. It's no longer a choice of either speech recognition automation or 100% human, and never the twain shall meet. Siri has dramatically increased customer expectations in terms of speech recognition, but most call centers' automated speech recognition engines (ASRs) still lag behind. And given that natural language processing is progressing in inches rather than by leaps and bounds, go for the hybrid model: improving your current ASR by adding pinpointed human intervention. Tools such as Spoken's Smart Interactive Voice Response (IVR) add pinpointed human assistance to IVR automation, which can improve automation rates by over 50%.
The last five years have brought many changes in customer behavior, thanks to speech-recognition-based tools such as Siri as well as readily available self-service through the web channel. And while customer queries have become more sophisticated, the call center is slow to adapt. It's time to reevaluate your call center best practices and take advantage of these new trends to upgrade, adapt and modernize. Your customers will thank you!
Think you know contact centers? Think again. This ain't your grandma's contact center.
As enterprises continue to hone in on improving the customer experience, the call center industry has been undergoing a much needed image overhaul. Unfortunately, reversing consumer perception of call centers is proving to be more difficult as previous negative experiences overshadow innovative new advancements. The reality is that call centers have become the cornerstone for customer service strategies and are becoming known as customer loyalty engines for enterprise. Below we’ll explain how current innovative call center technologies trump some of the closest-held consumer misconceptions about call centers.
Myth 1: Call center agents really don’t care about customer service
For many consumers, it’s a common misconception that the call center agent who handles their query doesn't care about providing quality customer service. On the contrary, because agents are on the front lines, business leaders acknowledge the business value and cost benefits attached to their role in improving the customer experience and make it a priority to foster agent success. In fact, the quality of customer care provided by call center agents is directly tied to the corporate culture and the reward system offered by the enterprise.
According to a recent study conducted by Deloitte Consulting, 62% of organizations view customer experience provided through contact centers as a competitive differentiator and are looking at various ways to further agent productivity. Coupled with the recent advancements in contact center technologies, agents are now armed with tools and real-time data to more effectively handle customer queries. Many companies also offer quality assurance programs to ensure even higher levels of service.
Myth 2: No one uses the phone anymore, anyway
There is no question that customer care channels are expanding beyond the phone to include SMS, social media, chat and smartphone applications. However, that doesn’t mean that phone-based call centers are on the decline just yet. In fact, Forrester Research reports that 69% of online consumers used the telephone to speak with a customer service agent. Additionally, Forrester Research reports that telephone customer service has the highest level of customer satisfaction compared to other online customer service channels. That being said, the industry is noting a strong preference among consumers for self-service; 72% of US online consumers prefer to use a company's Web site to get answers to their questions rather than contact companies via telephone or email. Taking this further, a recent survey by Aberdeen Group indicates that 65% of businesses use at least six touch points to engage with their customers.
The birth of "right channeling"
So while the call center isn't dead, it's transforming into a multichannel contact network. Smart companies are focused on responding to customers' needs via the customers' channel of choice and supporting them through the entire customer journey, regardless of channel. Multichannel technologies are helping enterprise call centers keep pace with new engagement strategies. Enterprises are increasingly looking for ways to incorporate chat, social media, texting, apps and web functionalities into their call center platforms. In fact, Deloitte Consulting found that 33% of contact centers already provide social media contact channels, and that number is expected to grow in 2014. So rather than proclaiming the death of the call center, many brands are embracing "right-channeling," or allowing customers to receive consistent service via their preferred channels.
Myth 3: Call centers are exclusively cost centers
Call centers have long been viewed within enterprise as a cost center only capable of improving the bottom line by trimming costs. However, that view omits one key function that call centers perform better than any other within the business: customer retention. Only the call center can turn a frustrated customer into a loyal one.
McKinsey analysts found that 70% of buying experiences are based on how customers feel they are being treated. On the flip side, Forrester found that 89% of consumers who experience poor service with a brand will leave for the competition.
Agent job satisfaction equals better customer service
And one of the best ways to ensure a positive customer experience is to improve the frontline agents' job satisfaction levels. According to a recent survey conducted by LiveOps, 50% of call center agents report increased productivity when they have access to “fully integrated social, multichannel desktops,” which translates to a more positive customer service and increased customer loyalty. The study also revealed that the incremental profit contributions from higher agent productivity ranged anywhere from $815K to $2M per year. In short, happy agents make for happy callers and an improved bottom line.
Because call center agents are communicating directly with customers, companies that emphasize customer service and agent job satisfaction achieve greater loyalty, better acquisition and increased spend. The research says it all: a positive experience with an organization’s call center has a direct impact on revenue.
Myth 4: Speech recognition doesn't work
In the age of Siri and IBM’s Watson, consumers have heightened expectations for automated speech recognition capabilities, which sometimes work against call centers. While Siri was built to handle a wide variety of tasks and commands, call center IVRs struggle to compete with Siri’s broader grammar and categorization capabilities.
In the call center, the key to accurate call routing rests in the ability to understand the reason for a call and identify the caller; therefore, it's critical that the speech recognition engine capture and interpret those caller utterances with near 100% accuracy. Advancements in speech recognition software are helping to bridge this gap. Enterprises are now able to compete with the likes of Siri using innovative hybrid solutions: provide human silent guides that work in the background to correct utterances that would otherwise fail (Find out more about how this works). The silent guides are able to supplement the automation and raise the speech recognition accuracy level closer to 90%, giving the callers a more Siri-like experience.
Myth 5: Remote agents cause security risks and decreased productivity
Security A common misconception of the remote agent model is that it creates a threat to private information. In actuality, virtual call centers can add security tools that ensure a higher level of security than is typically available in most brick and mortar call centers. Security measures such as intrusion detection, call recording storage, historical log files and PCI compliance audits are often included in virtual offerings. In fact, it’s much easier to lock down computers in a virtual environment at a comparatively low cost.
Also, new virtual products such as a Secure IVR allow a live agent to transfer a caller to a unique, masked PCI-compliant IVR to collect sensitive information in a secure manner and then transfer that data to back end systems, all without the agent ever coming in contact with credit card numbers or other sensitive data.
Productivity Additionally, people often fear that work from home agents might show lower productivity as a result of distractions or lack of workplace camaraderie. However, most virtual call centers actually report higher productivity because of the value the agents themselves place on the experience: working from home, eliminating the commute and working more flexible hours are seen as privileges. The virtual workforce tends to work hard to retain those benefits and, in turn, those happy agents offer better customer service. Organizations that implement remote agent teams report increased agent retention, reduced agent turnover and increased agent morale.
As we begin to examine the tip of the iceberg when it comes to the biggest call center misconceptions, it’s increasingly evident that times have changed and some long-held myths need to be exploded and demolished. New innovative technologies are equipping today’s contact center agents with the means to continue pushing the limits of exceptional customer service.
It's not easy running a contact center in the age of Siri. Customer expectations of speech recognition are more complex and higher than ever, and traditional speech rec IVRs can't keep up. What is a call center to do?
And many of those apps are voice-controlled apps, Siri being the most popular among voice recognition tools.
The smartphone experience with speech recognition technology has already changed user expectations of interactions with automation in the contact center. Your callers might be wondering, "If Siri can understand me, why can't you?" The Siri Effect is hitting call centers hard, and all IVRs and call flow designs should be under scrutiny.
The truth is that in the age of Siri, DTMF (Dual Tone Multi Frequency signaling, or touch tone response such as "Push 1 for sales, 2 for billing," etc.) doesn't cut it anymore. And customer care facilities should see this trend in user sophistication with respect to speech recognition as an opportunity rather than a curse.
The smackdown: IVR vs. Siri
While user expectations might be the same for a call center IVR as for a mobile speech-enabled app, the two have until now been vastly different in their goals and structures. Take a look at the structure, context and business expectation for each:
And due to changing user expectations, the contact center is experiencing a major shift in perceived value. This goes beyond asking callers to rate their satisfaction with a particular agent or a single call; these shifted expectations in terms of speech recognition interactions affect the entire corporate culture and the contact center's reason for being.
Paradigm shift: from cost center to loyalty engine
In the call center world, the old paradigm was to see the contact center exclusively as a cost center. This drove the quality metric of Average Handle Time, because the less amount of time spent on the phone with the customer, the lower the cost, right? This also drove organizations to create DTMF menu trees: who cares if the caller has to listen to three levels of menus and make six menu choices, as long as the forced-choice DTMF menu is easy for the organization to create? The organization didn't hesitate to make the caller do the work of navigating the DTMF IVR menu, because this is a cost center, anyway, right?
The paradigm that contact centers must adopt is the one that users have now come to expect: the contact center is a customer loyalty engine, not a cost center. Average Handle Time must be replaced by First Call Resolution, which is what the customer (not the organization) truly cares about. Users now expect accurate speech recognition, which is a little more work for the organization to create compared to a forced-choice DTMF menu, but it is much easier and more natural for the user to navigate. And speech recognition makes the technology do the work, not the caller. This is the new benchmark created by Siri: use technology to focus on a friction-free customer experience.
Why IVR speech recognition doesn't match the Siri experience
Let's take a look at how the Automated Speech Recognition engines (ASR) work in a typical speech recognition IVR call flow within the contact center. Let's say that Acme Widgets, Inc. has a speech recognition IVR for call routing, and the first prompt asks the caller "How may I help you?" The caller makes some sort of speech utterance as the Reason for Call (RFC), which the engine then parses compares to a language model of dynamic data based on hundreds or thousands of caller response utterances and converts to text for reference and improvement.
Based on the comparison of the converted utterance to the language knowledge base, the RFC is categorized into one of a series of buckets, such as sales, billing, customer service or tech support. Then the call is routed to the correct agent skill.
Where most contact center ASRs fail in comparison to Siri is between the language model and categorization. Because call center grammars are traditionally narrow, it's common for any unexpected utterance to fail because the developers didn't anticipate exactly what the caller might say. In this highly structured environment, one unclear syllable can make the difference between a smooth caller experience and a frustrating one.
How to match the Siri experience in the contact center
I'm not sure how other speech technologists have dealt with the rise in speech recognition expectations in the wake of Siri, but I'll share with you Spoken's solution to the issue. Since understanding the reason for call and identifying the caller are keys to accurate call routing, it's of vital importance that the speech recognition engine capture and interpret those caller utterances with near 100% accuracy.
However, within the structure of the contact center environment, even the very best ASRs out there only return about 50% accuracy for any one customer utterance. That means that even if the ASR understands nine digits of the phone number the caller utters, the system will still fail for lack of the tenth digit, and the caller will have a bad experience.
The Spoken approach is to support the technology with a hybrid solution: provide human Silent Guides that work in the background to correct those utterances that would otherwise fail. (If you're curious, find out more about how this works.) The Guides supplement the automation and bring the speech recognition accuracy level closer to 90%. If you're curious about the cost comparison, the quick answer is that because the Guides handles up to 10 simultaneous calls and reduce agent load, most customers experience a cost savings of about 15% over ASR alone.
What's really valuable is that the human safety net provides what pure ASRs can't: an additional layer of accuracy that improves speech recognition and keeps the experience closer to that of Siri.
Sophisticated speech automation for the win
Whatever approach organizations take to addressing the contact center in the age of Siri, the more sophisticated expectations of their customers can't be ignored. And while sites like GetHuman would suggest that what callers really want is a human agent, both Siri and studies have shown that users prefer automation over a live agent for simple tasks. And while 80% of callers will attempt to opt out or game the system when presented with a DTMF menu tree, over 90% will respond when asked an open-ended question such as, "How may I help you?"
Users are expecting more from speech-enabled IVRs than ever before. How is your organization addressing those heightened expectations?
There is nothing new that can be done to differentiate brands via IVRs.
Unless, of course, you're a Norwegian bank sponsoring a boys' choir. In that case, you can totally rock the IVR by having the boys' choir sing every prompt.
Really. Must be seen to be believed! Automation does not have to be boring. It can be downright moving.
And note this stunning statistic: in December, the telebank was called more than two million times. The population of Norway is 4.9 million. Think you can't distinguish your brand through your IVR? Think again.
Interestingly, most of these complaints are unique to IVRs. When describing human interactions at a brick and mortar store, for example, there may be complaints about wait times, but no one ever complains that he can't reach a human in a store or that the menus are too difficult to navigate. Humans are all around in real life--but can they solve your customer service problem?
The truth is that a well-designed IVR solution can actually address all three of these common customer service complaints and provide an extremely effective customer service experience. Let's address each issue individually:
Waiting too long on hold. Hold time is the bane of call centers and contact centers everywhere, and the balance of qualified staff to quantity of callers can be a difficult one to maintain. And yet, there are easy alternatives to forcing your callers to suffer through endless hold music.
First, consider offering self-service options, either in the IVR or online. For any industry call center, there is a list of the most commonly asked agent questions along with a procedural list highlighting how to correct them. If the issue has a common, standard solution with a proven success rate, there is no need to make callers wait on hold. Offer them the chance to self-serve, either in the IVR or online. Like the self-serve checkout lane at the grocery store, sometimes it's just faster and less frustrating if the customer does it herself.
Second, offer a callback. Within a well-designed IVR, when wait times pass a certain threshhold, the option to receive a callback is available. The process is simple: callers select the option for a callback and enter their phone number through DTMF or speech recognition while retaining their priority in the queue. The caller hangs up, goes about his life, and receives a call back from an agent when the queue clears.
Third, offer chat support. Chat support offers the immediacy of phone interaction with a built-in lag time that allows agents to multitask. Currently, chat queues tend to be shorter than phone queues, and customers can get near-immediate answers. While some still may prefer the IVR channel, deflecting those comfortable with chat support to that channel will ease the burden on the call center queue and reduce hold times for everyone.
Can't reach a human. We've written about this issue before. The issue isn't that the caller can't reach a human; the issue is that the robots aren't well-designed. A well-designed IVR will quickly collect enough information to identify the caller and promptly route the caller to a human agent. What is frustrating for callers is an IVR designed by engineers rather than customer service experts; what is logical to a technician may not be a logical call flow to a real human caller. Tip: if you want to know where your IVR isn't working, ask your front line agents. They will tell you exactly where the IVR is failing callers.
Navigating convoluted menus. There is no excuse for this one; IVR menus should be simple, direct and reflect the most common customer needs. A few basic rules:
No more than five choices. Humans can remember three or four choices easily, five with some difficulty. Giving more than that will frustrate callers.
No more than two choices on a submenu. If your submenu has more than two choices, you need to redesign your IVR.
More than one submenu. Having to navigate past even one submenu creates a barrier between the caller and the agent. Throwing two or three submenus in the call flow makes the experience more mazelike and less customer-friendly.
Use speech recognition. Pushing 1, 2 or 3 is fine for simple tasks, but it makes for a much better and more personal caller experience when the caller can simply say "I need help with billing" or "I'm having a problem with my printer."
When designing a customer response system, designers should keep in mind the most basic of human needs: callers simply want to be understood, helped and sent on their way. In many cases, this can be achieved with well-designed automation. In fact, studies have shown that for many routine tasks, customers prefer self-service to engaging in small talk with a live agent. Activating an effective speech recognition solution, designing and IVR with intuitive menu choices and offering callers callback and chat options address these common customer service issues as well as the human needs behind them.
The Wall Street Journal recently reported that supermarkets, the first to embrace self-service at point of service, are taking two very different paths. The Journal reports that while 200 Albertsons stores in the Southern and Western U.S. are phasing out the self-checkout systems and replacing them with human cashiers, the Kroger supermarket chain is continuing to embrace self-service, even to expanding to trials of a new scan-tunnel system that would automatically scan items from a conveyor belt.
To self-serve or not to self-serve, that is the question. And store owners seemed to have whittled the question down to a much simpler one: do supermarket patrons really prefer the personal touch or quick anonymity?
As with the answer to every other business questions, the answer is "it depends." On how well you know your customers and track their actual behavior. In this case, it depends on the type of supermarket patron and the type of shopping excursion are involved.
The basket excursion I don't know about you, but on most of my trips, I have no desire to bond with my supermarket checkout clerk. Several times a week, I am invariably grabbing one of those small baskets and squeezing the errand in between other activities that I actually enjoy doing and tend to approach the chore with a stoic pragmatism: get in, get the items on the list, and get out. I don't want to have a conversation with the butcher; I don't want to converse with other patrons; I don't even want to have an Excellent Customer Experience. I just want to buy my 10-15 items and then spend time doing something I actually value in life: interacting with my friends and family.
The shopping cart excursion However, that is only one kind of shopper one one type of shopping excursion. I'll admit that when I'm doing a monthly stocking-up trip, one that requires actually getting a shopping cart, filling to the brim and inevitably brings a three-digit total, the self-service checkout is not my first choice. In that case, the human checker saves considerable time and effort.Then again, that trip occurs once a month for me; the other, weekly or bi-weekly trips are all much better candidates for self-service.
Basket or shopping cart? One would hope that both Albertsons and Kroger's did studies to determine which type of shopper and excursion were more common, brought more income and inspired more brand loyalty before making their respective decisions to phase out and increase the level of self-service in their stores.
The lesson for IVR design
And we can learn a lesson here when designing IVRs. Most annoying IVRs do not fail because of the technology; they fail because of a flaw in the design logic. Just as a supermarket needs to consider its basket-versus-cart shoppers and their needs and values, companies should consider the following business logic before attempting to automate any process in the IVR:
What is the task most frequently requested by your customers?
What is the task your agents find the most boring and routine?
What are the top three reasons customers call your company?
What routine task are agents tired of repeating?
When your customers call after hours, what are they requesting?
When your customers call first thing in the morning, what are they requesting?
Understanding which tasks are vital to your customers and which are simply routine can and should direct your IVR design. It's the same as understanding which shoppers on which excursions want to simply get in and out and on with their lives versus which truly want that human touch. Remember that not all tasks fare well in automation: you must analyze your current customer actions in order to determine which tasks they will sing your praises for automating and which they will be grateful to speak to a human being over.
It's entirely too easy to mock badly-designed and badly-operated IVRs. But George Hrab does it entirely too well. Ah, if only "piloting the Space Shuttle" or "figuring out how to get more cheese onto America pizzas" were deep-down menu options, we might listen more carefully to those give layers of menu choices!
As someone who has been in on the painstaking process of figuring out, "Wait, if we ask 'Do you know your account number?', will they say 'Yes' or will they give the account number? Wait, no; we'd better be clear and say 'Please say your account number.'" I can testify that creating a nice, smooth IVR call flow that is short and sweet and that understands callers when they speak is a thing of beauty.
Bill Meisel of TMA Associates is the guest speaker revealing how the mobile phone speech interface has grown and changed the call center and IVR game as well as growth potential for change in the call center.
Bill Meisel, PhD, is publisher and editor of Speech Strategy News, co-organizer of the Mobile Voice Conference, Executive Director of the non-profit Applied Voice Input Output Society and a consultant on speech recognition technology markets.
The full screen capture will be posted shortly. In the meantime, enjoy the audio and slide deck: