SERTAINTY_SpeechTechsupp

							Produced by: Speech Technology Magazine

2

Introduction
now offers a solution to contact center managers to ensure quality, gather business intelligence, and ultimately serve the customer better. For those of you interested in applications serving call centers and customer service solutions, you will be happy to hear that Speech Technology Magazine will now be able to leverage the editorial, online, and infrastructure assets of its new sister publication, CRM magazine, as a result of the recent acquisition of AmComm, Inc. by Information Today. making, with our network of magazines, newsletters, Web sites, conferences, and books. The contact center is the natural overlap of Speech Technology Magazine and CRM magazine where we hope to extend the value chain of both publications and Web sites.

In this special section of Speech Technology Magazine, you will find an overview of the major applications for speech analytics within the enterprise environment provided by Datamonitor, followed by a concise discussion of the role of speech analytics in quality monitoring by SER Solutions. As with many data collection activities, the ability to analyze information is most often outpaced by the amount of data collected. In call centers this problem is exacerbated because audio data is unstructured, and until recently it was very difficult to structure the data and mine business value from the thousands of conversations taking place in a call center. Speech analytics

If you would like to view an online version of this section, Speech Technolog y Magazine has made this section available in PDF form by download through our Web site, speechtechmag.com. And because these topics are complex and deserve more Our goal is to provide you, than cursory consideration, I our reader, with end-to-end encourage you to get more content, from business information from both SER’s technology solutions through and Datamonitor’s Web sites. strategic executive decisionBob Fernekees Group Publisher BFernekees@Infotoday.com Information Today, Inc.

Supplement to Speech Technology Magazine - September/October 2006

Quality Monitoring and Business Intelligence with Speech Analytics
Section 1

He Said, She Said

3

By Ri Pierce-Grove The volume of archived audio communications is growing worldwide. As governments and enterprises look for ways to manipulate and mine those archives, speech analytics is gaining ground, moving from an emerging to an applied technology. In essence, speech analytics applications allow users to organize recorded calls for content. There are two major threads of technology behind speech analytics — speechto-text and phoneme-based — but both function as search engines for audio, responding to queries with lists of calls in which a given interaction took place. Although concerns about accuracy still exist, a list generated by a speech analytics application of calls in which variants of the words “cancel my account” were spoken is a considerably more valuable guide to customer interaction than random sampling. These search engines can be used to meet two distinct needs of the customer-centric enterprise: quality monitoring and isolation of customer trends. Or, to put it another way, they can be deployed

to answer two questions: “What are our agents doing?” and “What are our customers trying to tell us?” Several economic trends have come together to make speech analytics a logical extension of the customer-centric enterprise. As the cost of data storage has dropped and technology has improved, the percentage of calls being routinely recorded has increased. In the background, the

development of CRM as a philosophy and a technology means that firms are increasingly inclined to see customer interactions as a set of data to be organized and analyzed in order to better serve their customer base. Analytical and operational CRM has been providing significant support, but until recently, the technological obstacles to analyzing audio meant that firms were unable to perform the same actions with recorded calls.

250

200

150

100

50

0

Supplement to Speech Technology Magazine - September/October 2006

4
Revenue Total Growth 2005 50 2006 66 32%

Quality Monitoring and Business Intelligence with Speech Analytics
Section 2

He Said, She Said

Table 1: Global revenues for speech analytics, 2005-2010
2007 86 30% 2008 112 30% 2009 159 42% 2010 218 37% CARG 34%

Speech analytics is therefore a potentially valuable tool for enterprises interested in converting their growing archives of recorded interactions into data assets. In the last three years, an increasing number of firms have responded to this need by entering the commercial sector with a speech analytics offering. Call recording providers, intelligence and security firms, speech technology vendors and stand-alone analytics providers are all proffering solutions. As of March 2006, the majority of deployments were in a pilot phase; and speech analytics has gained ground since.

automating the search for calls which merit their attention. By searching for variants of the phrase “close my account,” for example, users can rapidly identify not only callers who actually closed their accounts, but also callers who simply threatened to close their accounts in reaction to poor customer care. Speech analytics can be used to identify agents who diverge from script to find examples of both inappropriate and neglected upsell and to immediately locate calls which may expose a firm to liability.

call volume can use speech analytics to find out why its customers are calling them so often and identify calls which could be automated. A telecom company concerned about churn can use speech analytics to find out if callers are mentioning its competitors. In general, speech analytics offers firms a way to get in touch with a much larger set of information on their customers than a focus group can offer.

Adoption: Incentives & Practice
The key factors which suit a firm to early adoption of speech analytics are large call volumes, existing investment in call recording, strong potential for churn, and concerns about liability and regulatory compliance. Heavily regulated industries like banking already have large volumes of recorded audio, which gives them an existing data asset to exploit. In addition, it is imperative for them to be able to demonstrate that their agents are consistently obeying regulatory requirements; and they are heavily concerned with questions of fraud. Speech analytics

Business Intelligence
The opportunities for business intelligence or identifying customer trends are more varied. Marketers can use speech analytics to find out whether a given promotion is generating calls. An insurance company interested in reducing its

Quality Monitoring
Speech analytics can help quality monitoring personnel spend their time more efficiently, by

Supplement to Speech Technology Magazine - September/October 2006

Quality Monitoring and Business Intelligence with Speech Analytics
Section 3

He Said, She Said

5

can help address both of those issues. Firms which offer highly substitutable products and are concerned about churn — like telecom providers or mortgage brokers — also have advantages to reap from speech analytics. They need to distinguish themselves in customer service and be able to react rapidly to their competition; and speech analytics can help speed up their reaction time. Contact center outsourcers have a high volume of calls by the nature of their business and are

looking for ways to differentiate themselves and stave off commodification. For them, speech analytics offers a welcome opportunity to assuage potential customers’ concerns about loss of control and security with regular reports on both agent behavior and caller trends. A given speech analytics solution may be suitable for both quality management and business intelligence purposes, but different sets of people will be using it in each case. Deployment of speech analytics for quality monitoring (QM) can deliver value with comparatively

little cultural change for a firm, since QM is generally a distinct departmental function. In order to use speech analytics effectively for business intelligence, however, firms should give serious attention to how the solution will be used within the enterprise. As with any business intelligence tool, its ROI depends on its exercise. A valuable tool used ineffectively runs the risk of being another expensive, but useless add-on. Therefore, as enterprises move towards speech analytics, they must have a clear vision of what they plan to do with it internally. Will its use be concentrated in the hands of a few power users who have heavy data management and delivery needs and a limited relationship to the rest of the enterprise, or will it be used more broadly by line-of-business managers? How will the insights gleaned be disseminated within the organization, and what decisions will they inform? Speech analytics is a logical extension of the customer-centric enterprise, and this is a critical moment for this market. Both clients and vendors have the opportunity to seize strategic advantage before the market shakes itself out.

Supplement to Speech Technology Magazine - September/October 2006

6
Mark Clark Executive Director Product Management

A Faster, More Effective Way to Ensure Quality in the Contact Center
Section 1

Ensuring quality in service delivery is a top priority with every contact center. One of the most critical drivers of service quality is effective coaching. A common frustration of supervisors is the inability to spend enough time coaching agents. In a recent vendor survey, 78 percent of supervisors said they spend less than 20 minutes per day coaching each agent. A major constraint is the amount of time spent in Quality Assurance monitoring and evaluations. Although virtually all large contact centers now have an automated quality monitoring system, the process is still slow and inherently inefficient. Benchmark Portal found that the average supervisor spends 3.27 hours per month evaluating each agent. Following are the leading difficulties experienced by contact center supervisors. Small sample – Even the best contact centers only review 1-3% of the total calls handled. This small sample may not be statistically representative of the calls actually handled. Agents can quite

convincingly protest that the calls reviewed were not reflective of the broad majority of calls actually processed. Conversely, an agent handles many calls well; however, he/she has intermittent challenges with a certain type of transaction. A small sampling may not capture any examples of those particular calls so that additional training and/or coaching can be provided. Searching for coachable calls – Supervisors have to listen to many more calls than are actually scored as they search for a “coachable moment.”

Lack of objectivity – Agents are rated on subjective criteria like tone, professionalism, clarity, and courtesy as well as objective KPI’s. Different supervisors and monitors will have different evaluation standards for these soft skills. Valuable input is unused – If only a tiny portion of the calls handled is later reviewed, the company will miss important customer insights. These could lead to product and service improvements, alerts to competitive activities, feedback on the effectiveness of marketing campaigns, or awareness of potentially serious quality issues.

Assessing Which Calls Have Value

Supplement to Speech Technology Magazine - September/October 2006

SER Advertorial

A Faster, More Effective Way to Ensure Quality in the Contact Center
Section 2

7

Liability exposures – Contact centers must comply with a multitude of federal and state requirements, particularly if the contact center is engaged in selling (including up-selling and cross selling), collections, or healthcare. State and federal regulations such as requirements of the Federal Trade Commission for Express Verifiable Authorization (EVA) demand that vendors maintain a record of purchases and customer details, and that specific disclaimers be read to the customer.

Finding coachable calls is easy – All calls are readily accessible through an easy to use GUI. Supervisors can search and retrieve calls based on pre-specified criteria or by using search tools similar to Internet search engines. SERTAINTY uses a phonetics-based search engine. This is more flexible and forgiving than dictionary based speech systems. The result is less time listening to routine calls and more time for coaching. Greater objectivity in defining soft skills – Supervisors can compare rated calls in calibration sessions to improve consistency. Agents will feel more confident that the evaluation process is fair and objective. Gather important business intelligence – Since SERTAINTY can search all of your recorded calls, the contact center can become a valuable source of business and market intelligence. For example, your company is launching a new service and you are kicking it off with an exciting trial offer. Simply include the words “trial offer,” the brand name, and other related words and phrases and set up a bucket in the rules manager. Using the built in reports, you can graph the frequency of mentions to determine the level of impact.

The Promise of Speech
SERTAINTY by SER Solutions, Inc. addresses all of these issues through the use of modern speech technology. Let’s take them one at a time: All calls are reviewed rather than a small sample – SERTAINTY takes the audio output from virtually any contact center’s recording system and listens to it for you. Unlike a human monitor, the system can review all calls, searching for key concepts and phrases you deem of relevance to your business. You can set up your own categorization system to file calls for later retrieval and analysis. Now, all of an agent’s calls are subject to review.

Assure compliance – Say you are in the collections business and your agents are required to state a “miniMiranda” at the beginning of each collection call. Or your company sells insurance, investments, banking services, wireless services, or virtually any product or service. An alert agent can turn a general service inquiry into a sales opportunity but then certain disclosures need to be stated (and certain words and phrases should never be said). SERTAINTY will automatically check that the correct scripts were read, and even issue alerts to supervisors if they are not stated at an acceptable level of precision. The result – Improved quality control in the contact center Speech technology has been around for a couple of decades. Early iterations were problematic. Now speech is truly ready for “prime time” and finding its way into the contact center for self-service and quality assurance. SERTAINTY can examine a virtually unlimited number of calls – something it would take an army of supervisors to do. To the contact center, this means greater objectivity, accuracy, consistency, and improved agent morale. For your customers, it simply means better service.

SER Advertorial

Supplement to Speech Technology Magazine - September/October 2006

8
Speech Technology Magazine www.speechtechmag.com 2628 Wilhite Court, 100 Lexington, KY 40503 P: 859-278-2223 Toll Free: 877-993-9767 Fax: 859-278-7364

Contact Info
Information Today, Inc. www.infotoday.com 143 Old Marlton Pike Medford, New Jersey 08055 P: 609-654-6266 Fax: 609-654-4309 SER Solutions, Inc. www.ser.com 45925 Horseshoe Drive Suite 150 Dulles, VA 20166 Email: info@ser.com
(Please use “Speech” in the Subject Line)

Supplement to Speech Technology Magazine - September/October 2006