The Problems Are Solved: Quality, Integration, and Security

By Mark Swanson

This is Part 3 of 3 in a series about How VocalQ Has Already Changed Your Business, The Personality of Speech Recognition, and The Problems Are Solved: Quality, Integration, and Security

It is no secret that the amount of data in our world has been growing exponentially. The term “Big Data” has been coined referring to the ability to analyze very large data sets.  According to McKinsey, Big Data “will become a key basis of competition, underpinning new waves of productivity growth, innovation, and consumer surplus.” Big Data provides the ability to take all this data we have been generating and make better decisions based on it.  A recent study by IBM identified that one in three business leaders make decisions without the information they need and half don’t have access to the information they need to do their jobs. That has significant competitive implications. Once the data is extracted and analyzed, a whole world of possibilities opens up.

SIRI is just a crack in the door.

Imagine scanning tens of thousands of voice conversations both in and outside of your company, recognizing what is said in those conversations, and then being able to map and measure both the quantity and quality of the conversations that took place.  We are doing it on the Web, why not on the phone? According to Forrester, over 73% of businesses are spending a combined almost $1billion on Web analytics technologies. Imagine being able to capture and rate each customer interaction, and be able to select any conversation and play it back, offering either encouragement or praise for that employee.  We agree with Norman Winarsky’s comments, not specifically for voice recognition, but rather for the voice recognition combined with voice analytics.

In addition to leveraging data, performing speech recognition and analytics in the Cloud has solved other obstacles that have been holding the technology back: audio quality, complex integration and security issues.


The quality of voice signal at the network core is as high as it is going to be while traveling inside the telecommunications network.  Capturing uncompressed audio signals at the core of the network means you eliminate transcoding issues, clipping and other effects that can disrupt voice recognition.  In addition, many of today’s Cloud Voice Networks are operating with the wide-band audio (G.722) standard which allows the voice to be captured with pristine quality.


Another issue that the Cloud solves is the ability to capture and integrate multiple collection points into one central repository.  Multi-site organizations have greater costs and higher integration risks than single locations.  The costs and risks have held back the deployment of these technologies.  Centralizing the collection and analysis of voice conversations allows for a much simpler deployment and vastly less integration.  In addition, the ability to connect to large processors to crunch the data and deliver analytics provides for a faster experience.


Deploying voice recording and analytics technology on premises opens up another can of worms.  Establishing and assuring consistent security across all points of control becomes a very difficult task for a typical IT department.   Recording in the Cloud greatly simplifies the security and risk management of operations.   In the Cloud, a policy can be enforced and checked on by the customer.  Many Cloud vendors have established security and privacy policies that are verified by third parties.  In addition, most have backup systems that are used regularly.   Also, the ability to monitor and lock out “insiders” is of great benefit to companies.

The widespread adoption of Cloud technologies is providing a significant boost to call recording.  Are you prepared for the shift?  How do you plan on implementing these changes into your business?  Share your thoughts below.


The Personality of Speech Recognition

By Mark Swanson

This is Part 2 of 3 in a series about How VocalQ Has Already Changed Your Business, The Personality of Speech Recognition, and The Problems Are Solved: Quality, Integration, and Security

Speech recognition and processing technologies are on the verge of an explosion in adoption for both consumers and businesses; 2011 saw dozens of new voice enabled applications launched for Android, iOS and Windows.  One of the most visible was launched in October 2011 when Apple announced the iPhone 4GS and its natural language voice control system called SIRI.  This most popular new feature on the iPhone uses results of over 40 years of research funded by DARPA and organizations and universities across the United States. SIRI co-founder Norman Winarsky was not shy about sharing his thoughts on how SIRI will not only change computing, but the entire world.  “The PAL (personal assistant software) will get things done, and this is only the tip of the iceberg.

We’re talking another technology revolution; A new computing paradigm shift.

While that statement may be hyperbole, something has changed to generate interest in using voice recognition.    That something is integration of voice recognition with the analytics capabilities of Cloud technologies.

Analytics and the Web go hand-in-hand.  The ability to analyze interactions with customers is one of the primary reasons for the rapid adoption of ecommerce.  The enormous popularity of Amazon’s recommendation engine is a testament to this.  Web analytics allow businesses to measure, collect, analyze and report on Internet data for purposes of understanding and optimizing Web usage. Web analytics not only help companies measure Web traffic, but help businesses listen to customers through market research, measure advertising campaigns, determine popularity trends and analyze what works and what doesn’t – both in the aggregate and individually.  On-site analytics measure customer interactions on your Web site.  This includes drivers and conversations in a commercial context – what’s being searched for, what is being purchased, what customers are saying.

The reason SIRI works so well is that it uses Cloud-based technologies to process the context of the spoken language to continuously learn, not only what is said, but also what is meant.   This allows SIRI to have a personality.  And just like a human who fails to understand what you mean, it can fail graciously.  It will interject humor into the conversation, as well as ask clarifying questions.   For example, when one of my co-workers, in a moment of frustration, asked SIRI “Why am I such an idiot?” it responded with, “I have been asking myself that lately.”  When SIRI can’t find an answer to an ambiguous question it can search the Cloud and respond with humor or another question.  SIRI is leveraging a concept called “Big Data.”

Up Next: Part 3 – The Problems Are Solved: Quality, Integration, and Security

How VocalQ Has Already Changed Your Business

By Mark Swanson

This is Part 1 of 3 in a series about How VocalQ Has Already Changed Your Business, The Personality of Speech Recognition, and The Problems Are Solved: Quality, Integration, and Security

The axiom in the IT industry is that a seismic shift in the “computing platform” occurs every decade or so.  We saw it in the move from the mainframe to the mini-computer in the ‘70s, from the mini-computer to the PC in the ‘80s, the PC to Client Server in the ‘90s and to the Cloud in the 2000s.  We believe that the next shift in computing has already begun.   It has to do with how we interact with computers and how they understand that interaction.  We are now at the cusp of combining true voice recognition with the ability to analyze the “Big Data” generated by these conversations to provide immediate insights into what is happening behind the electronic veil.  If you are a business manager, you should pay attention to this emerging trend.

Missing this shift could mean going out of business.

Despite fifty years of technology advances, conversing with your computer has remained in the realm of science fiction.  If you are over 40, you probably remember listening to the tranquil voice of HAL in 2001 Space Odyssey or watching Spock bark orders to the unnamed Duotronic Computer on the Starship Enterprise.  In our mainstream culture, this future vision has yet to pan out – in fact as a culture we have gotten much more adept at using our thumbs than talking to a computer!

Despite our lack of awareness, our conversations are already being recorded, analyzed and tracked.  A monolithic network of computing power owned by the Government, called Echelon, is listening to and recording phone conversations.  Echelon provides a secret cadre of snoopers the ability to monitor and analyze millions of conversations.  And, these are not just military or terrorist calls.  In 2005, we learned that the NSA was wiretapping civilian conversations with Echelon.   To this day, our government continues the policy of secretly tapping into our conversations and recognizing what you say.

You are also being understood.

Improvements in technology are turning speech recognition into speech understanding.  The Defense Advanced Research Projects Agency (DARPA) has been working on a project for recognizing speech and understanding its meaning since 2005.  A project called GALE uses neural networks and statistical modeling technologies to “absorb, translate, analyze, and interpret huge volumes of speech and text in multiple languages.”

Commercial applications have lagged behind the government, but are now emerging as viable options.  After a gestation of several decades, commercial speech recognition technologies popped onto the scene in the mid 1990’s.  Apple and Microsoft started embedding speech recognition into their operating systems and Dragon Systems unveiled its software that recognized normal human speech.  These applications were prone to errors and problems with the technology caused it to fade from the market, however, something has changed in the past couple years.

Up Next: The Personality of Speech Recognition

4 Key Benefits of the New VocalQ℠

Used effectively, the VocalQ framework is a strategic differentiator and delivers considerable benefits.  Speech recognition and analysis has come along way in the last decade and the technology is poised to accelerate in the next.  The day is not too far off when speech recognition and analytics will morph from improving customer interactions to being the customer interaction.  Our goal with VocalQis to help companies realize the benefits:

Reduced Costs  – Companies gain valuable insight into the inefficiencies of their operations.  Speech analytics help to decrease call times by training agents to be more effective in harvesting information from customers.  Currently, many businesses measure quantitative call data like; how many calls were received, how many calls were abandoned, how many rings it takes to answer an inbound call, etc. This data only answers “what” communication is happening within your business. VocalQ measures the context and content of your communications.

Better Customer Retention – Cash spent on retaining customers through improving customer service is far less than that needed to gain a new customer.  In addition to measuring quantitative and descriptive data, VocalQ goes a step further by recording and alerting on the quality of interactions. With VocalQ, What's Your VocalQ?calls are sampled with call recording and the dialog is benchmarked.  Things like customers having to repeat themselves and correct responses to questions can be measured.  Furthermore, the recorded conversations can be compared against customer surveys so supervisors can discover why an interaction went well or went awry.

Cross Selling – Speech Analytics identify opportunities to increase sales conversion rates, both for new customers and up-sell and cross-sell opportunities.

Reduced Risks – Using speech analytics enables a business to quickly identify potential regulatory failures or locate poorly performing agents.  Rather than relying on gut feeling, you “hear” the voice of the customer, analyze the interaction in real-time, and make better strategic decisions about your company.  The ability to take immediate action on “voice alerts” is another benefit of VocalQ.  Supervisors can identify critical words and phrases in conversations and receive real-time alerts to remedy the situation now rather than later.

What other benefits have you experienced with the New VocalQ?

A New Framework for Evaluating Customer Interactions

Today we propose a new framework for analyzing and improving customer interactions within your organization.  Rather than rely on quantitative and descriptive measurements, such as rings to answer, we suggest shifting the focus to the quality of interaction and prescriptive measurements.  In this framework, calls are sampled with call recording, and the dialog is benchmarked.  Things like: customers having to repeat themselves, and correct responses to questions, can be measured.

What is Your #VocalQ?

Recorded conversations can be compared against customer surveys so supervisors can discover why an interaction went well or went awry.  Supervisors can listen for a tone that is professional and empathetic.  A qualitative score is determined and the best conversations then saved for future customer service training.   This framework for measuring customer interactions – both in a descriptive and prescriptive manner – is outlined below:

Measurement Criteria



Prescriptive Measuring metrics that directly effect customer satisfaction and taking action on them.

Example:  Average wait times exceeding prescribed limit

Action: Automatically put more agents in the queue

Monitoring actual conversations in the customer engagement and using voice triggers

Example: Recognizing words that indicate positive experience and saving conversation

Action: Measuring other agents against the positive interaction

Descriptive Data that describes the details about the call

Example: Call Detail Logs (answer time, time of call, duration, etc.)

Data or recordings that describe the details about customer conversations

Example: Measuring how long it took the customer to get their issue resolved.

An example of using this framework:

Record the first request of customers, and then capture the response of the service representative.  The reaction of a representative to a hostile customer is often key to defusing a tough situation.  Phrases like “I am sorry a mistake was made” versus “I am sorry I made a mistake” can steer the conversation in a different direction.   The context is what is important and prescriptive/qualitative data about what was said helps improve the engagement.

The ability to take immediate action on words that we call “voice triggers” is another benefit of this framework:

Supervisors can identify critical words and phrases in conversations.  Phrases such as “system outage” or “I am down” are used to trigger text messages to supervisors who can automatically add customer service agents into the queue.  The ability to address problems immediately improves the processes and procedures at the frontline that drive the business.

What is your Vocal Q?

Do you know if your voice interactions are consistent with what you want your brand to represent?

In today’s highly competitive market, every customer interaction represents your brand. This not only includes interactions with your contact center, but also every employee who engages with your customer.  Today we have tools to help measure customer engagements and evaluate the context of the interaction.  We have the ability to hear the “voice of the customer” and improve the interactions we have with them.

Poor customer interactions are a huge problem in organizations.  Nearly 70% of consumers said they had ended a relationship due to poor customer service alone.

What is the cause-and-effect of this?

The root causes of poor service are:

  • Being trapped in automated self-service
  • Being forced to wait too long for service
  • Customers having to repeat themselves
  • Representatives that lack the skills to answer customers’ inquiries

The vast majority of organizations are paying attention to the wrong metrics.  Most organizations are spending their efforts on quantitative data like queue holding times and speed to answer, or worse yet – relying on short surveys that are viewed as intrusive and rarely answered.  If firms want to ensure positive customer experiences whenever customers interact with the company, they must focus on qualitative data like:

  • “How long did it take my customer to get to the person who could solve their problem?”
  • “Did the representative(s) ask the customer to repeat themselves?”
  • “Did the representative have the skills to answer the question correctly?”

Without quality, quantity means nothing.

Organizations are not properly analyzing the data surrounding customer interactions – they are being descriptive when they should be prescriptive.  A prescriptive analysis examines customer interactions in order to determine what should be, rather than what IS.  In a customer experience context, this can be more useful to decision makers than basic phone log data.  A descriptive analysis, on the other hand, seeks only to measure and explain what is, rather than what should be.  A descriptive analysis describes reality without an opinion; it is a guide for future action, but does not directly advocate action.  If firms want to improve their customer interactions, they need to be prescriptive and link the data to actions that will directly impact a positive customer experience.

Are you descriptive or prescriptive?

‘Telephone’ was a Fun Childhood Game, but Do You Really Want to Play it in Your Business?

Do your reports match reality? Are your decisions based on reality or secondhand information?

‘Telephone’ is a game in which one person whispers a message to another, which is passed through a line of people until the last player announces the message to the entire group.  It is often regarded as a metaphor for cumulative error, especially the inaccuracies as rumors or gossip spread, or, more generally, for the unreliability of human recollection.

There’s no doubt this game is amusing as a child, but how comical would the inaccuracy of human recollection become if you played it in your business?  How much laughter would you still hear if you discovered one of your previously ambitious account executives could no longer close new business?

Would you giggle if you were no longer profitable?

In this fictitious game of Telephone, casual chatter about the latest opportunities circulate around the sales community over the course of the week.  The seemingly innocent conversations cause an ambitious senior account executive to feel the effects of miscommunication around the office.  Adam is a consistently high performer, but when his boss, Gary, hears the misshapen truth of Adam’s latest lead, tension arises:

Adam: “I’ve met with the client several times but they’re a bit tough.  I have another meeting tomorrow and am hoping to close the deal.”

Bill: “He’s met with the client several times and they’re tough.  He has another meeting and might close the deal.”

Christine: “He’s met with the client several times and they’re really difficult to deal with.  He might close the deal.”

Diane: “He’s met with that impossible client multiple times but he said he hasn’t closed the deal yet.”

Eileen: “He said he’s met with that impossible client but he still hasn’t closed the deal.”

Frank: “He said that client is impossible and he can’t close the deal.”

Gary: “Adam, are you calling our clients impossible?  If you can’t close the deal I can give it to someone else.”

Which company are you?  Which company do you want to be?

To Adam’s surprise, the perseverance and positive attitude he shared with a coworker earlier in the week had inaccurately passed through a line of people – just as in the Telephone game.  The snowball of miscommunication created a distorted perception of Adam because the inaccuracies of human recollection are unfortunately inevitable.  Adam was advertised as the exact opposite of determined; and in this case, there’s no evidence to support what Adam really said.

While this is a fictional example for our purposes, situations like this are real.  They happen everyday both internally and externally, through all channels of communication: in-person, through email, and on our business lines.  With all those channels of communication, how easily can you access the necessary evidence to fairly dispute customer complaints?

Just as your CEO wants the knowledge to accurately address your largest client’s concerns, you need this knowledge to moderately settle even the smallest discrepancies.  Quantitative data is great for good companies, but qualitative, context data turns the good company into a great company.  Take Company A and Company B, for example:

Company A knows how many phone calls they receive in the office and who is calling.

Company B knows how many phone calls they receive in the office, who is calling and on what device, who they want to speak with and what they’re saying, who they spoke with and what they said before, when a dispute is escalating, what the dispute is about, and where the dispute originated from.

Which company do you want to be?