Skip to content

Speech Analytics Is Starting to Make a Difference

Speech Analytics Is Starting to Make a Difference

With skilled analysts behind them, these solutions can enhance the customer journey.

By Donna Fluss

  Printer Friendly Format       View this document on the publisher’s website.

Post-call speech analytics is ready for prime time and all types of users. Real-time speech analytics is an emerging solution that is highly compelling due to its potential. Although the underlying technology for these two solutions may be the same, their uses are totally different. Post-call speech analytics is a strategic enterprise application that companies should use to rapidly identify operational, procedural, technical, and staff-related issues, as well as pinpoint new revenue opportunities. Real-time speech analytics is a tactical application designed to alter the outcome of phone conversations while the caller is still on the line. This makes it a contact center tool that can be used to reduce risk and minimize bad customer experiences. As a result, it also becomes an effective coaching tool for agents.

Post-Call Speech Analytics Nears Maturity

Speech analytics entered the commercial market in 2004, and its use has been growing rapidly due to its unique and highly compelling value proposition. Speech analytics remains the only application that can structure phone conversations and find insights and trends. The automation component of these solutions is the easy part of the process; the enterprise challenge has been figuring out how to apply the findings.

Speech analytics solutions are sophisticated tools that require management by highly trained people to realize the expected benefit and return on investment. However, contact centers, the primary users of speech analytics, prefer solutions that are highly intuitive and can be used by supervisors. This is the main reason many speech analytics implementations are not delivering the expected benefits. These solutions require ongoing care and modification—tuning, searching, and filtering—to deliver targeted and effective findings. Once trends and insights are identified, companies need a way to share this information on a timely basis, and a vehicle for driving change. Speech analytics is an enterprise business intelligence tool whose findings should be incorporated into a formal change management process. This tool and process should be overseen by skilled business analysts empowered to drive change throughout the enterprise.

Not All Solutions Are Created Equal

There is a misconception in the market that most speech analytics solutions offer similar capabilities. This is a result of the fact that almost all speech analytics vendors make similar claims about their solutions’ capabilities. First-time buyers do not typically have the expertise or knowledge to distinguish between the offerings. Solutions designed to spot key words and phrases, which are the most common in the market, cannot perform a forensic analysis and identify new trends, for example. Speech analytics packages built to address specific business issues and that come with predefined lexicons (libraries), searches, reports, dashboards, and key performance indicators are very different from solutions that come with a blank canvas where users have to build everything themselves. DMG cautions prospects to carefully evaluate the various solutions when making a selection, and keep in mind that if you are getting it for free, there is probably a good reason why.

Speech Analytics Reduces Risk

From the beginning, speech analytics has been used to measure agent script adherence—making sure that agents say what they are supposed to at the right time during a conversation and that they are not saying inappropriate things to callers. Script adherence remains an important use of speech analytics; this capability has a proven payback for companies that set it up properly and maintain it. But as governments and other agencies in countries around the world have introduced regulations to control the handling of sensitive customer credit card information, debt collections, sales, the calling of mobile phones, etc., speech analytics has become a valuable tool for proving that a company is in compliance. Companies are using speech analytics to help them adhere to the Payment Card Industry Data Security Standard and to improve their staff’s compliance with the proliferation of debt collection regulations. Real-time speech analytics is also beginning to surface as an important new capability for outbound solution providers who need to demonstrate compliance with the Telephone Consumer Protection Act.

QA Gets an Overhaul

Companies have been doing quality assurance (QA) the same way since this technology was introduced close to 40 years ago. Vendors initially introduced and marketed these tools as a way to improve service quality, and they did not sell. Only after the vendors tied quality improvements to a reduction in call average handle time did adoption pick up. It became clear that companies would invest in contact center solutions that improved quality and the overall customer experience as long as they also improved productivity. So, although traditional QA is a manual process enabled by technology, companies invested in these solutions because they delivered significant quantifiable benefits.

Analytics-enabled QA takes this to a new level. Speech (and text) analytics can be used to identify a variety of calls (and emails, social media interactions, chats, etc.) where agents do not follow departmental policies and guidelines. As long as a company can build a rule to check and look for certain things, speech analytics can find and monitor it. There are still many things that speech analytics cannot check for and catch. While speech analytics can identify if an agent properly delivered the mini-Miranda or verified a customer at the right time, it cannot determine if the answer given is correct or not. However, considering that most companies only check three to five (and possibly 10) calls per agent per month, or just 1 percent to 3 percent of all calls, applying speech analytics to 100 percent of calls improves the odds of identifying behaviors that need to be changed, even if it cannot catch everything. And, as time passes, more sophisticated solutions are getting better at catching more things. So, while analytics-enabled QA is not perfect, it is better than the traditional way of doing quality assurance, which does not yield statistically relevant results for most companies that use it.

It’s All About the Customer Journey

Companies are finally building multichannel (or omnichannel) and cross-channel servicing environments. Delivering a true multichannel contact center experience is very different from having separate groups of agents handle various nonintegrated media. Customers do not want to begin a transaction in one channel and start over when they move to a different channel. Customers rightfully expect a seamlessly integrated sales and service experience, whether they call, email, chat, or send a comment via social media.

Companies also need customer experience analytics to measure all “touches” in the customer journey, and speech analytics vendors have jumped at the opportunity to deliver packaged solutions. Vendors have started to offer integrated speech, text, and desktop analytics. Other vendors are combining speech analytics with surveying solutions to provide an internal and external view of customer satisfaction. DMG expects to see substantially more investments in these packaged multichannel analytics solutions during the next few years.

What to Expect in the Future

The speech analytics market has come a long way in a short time, and a great deal more is expected from it. Real-time speech analytics is in its infancy, but its potential is great, as it gives companies a new way of looking at and interacting with customers. More companies are going to integrate speech analytics with real-time guidance solutions to transform the way their staff handles customers. Speech, text, and desktop analytics will be integrated with predictive analytics solutions, and the output will be used to feed real-time guidance applications. More vendors are going to build customer experience analytics solutions that can capture and analyze customers’ behavior throughout their journey, regardless of where they start and end. It’s clear that speech analytics is useful on a stand-alone basis, but its value increases as it is integrated with other high-value applications and processes.