Speech Technology Magazine

Webinar -  Everybody's Talking: The Path to Conversational Customer Care

Register now for this September 26, 2017 Event!

Unify Communications Across Channels

Semantic parsing can mine a treasure trove of data
By Aaron Fisher - Posted Mar 1, 2011
Page1 of 1
Bookmark and Share

As Web and mobile devices skyrocket in popularity, consumers are using these channels to develop creative ways to express their needs, likes, and dislikes. That trend has created some challenges for organizations, which often find themselves managing channels separately. The result is a lack of cross-channel continuity and a disjointed user experience.

For many organizations, the issue is how to develop a cross-channel strategy that would unify communications. One workable solution would be to integrate speech technologies and techniques used in speech recognition. The goal is a seamless customer experience.
End users often give indirect feedback to organizations through social media and communities.

Indeed those channels are incredible sources of information, but they also present daunting challenges. The biggest is the overwhelming amount of data generated. Using a technique called semantic parsing, organizations could mine this treasure trove of data to identify the overall sentiment of their customer base, as well as areas for communication opportunities.

Semantic parsing is commonly used in natural language speech applications. At a high level, semantic parsing lets you derive a meaning, or multiple meanings, from text strings. Because of the large amount of textual data created through social media and communities, the technique is a perfect fit. Using semantic parsing would enable organizations to identify posts that deserve an immediate response, either automated or from a human.

Additionally, semantic parsing can be utilized to interpret the meaning of text messages. Texting is not widely used for business-to-consumer communications because of the structure needed to function in this environment.

By using a semantic parsing engine as a free-format text interpretation engine, however, organizations would be able to provide real-time self-service through this channel. The channel would be opened up for full-scale service application or it would be used as a transition into other channels, such as live agent and Web, that would best meet a customer’s needs.

Mobile applications continue to gain traction as a way to enable consumers to interact with businesses. These applications are growing more complex, yet some barriers must be removed for this channel to reach its potential. Integrating speech into these apps is a good start.

For example, today if a consumer using a mobile app encounters something that is not available to him, he must abandon the mobile session to pursue another channel. By simply integrating speech recognition into mobile, users would be able to speak into their phone to leverage a different channel when the need arises. This cohesive user interface would reduce the need for users to employ multiple individual channels to complete a transaction.

Virtual assistance (VA) applications are another way to integrate speech into a mobile Web application. Using speech’s natural language in conjunction with the VA’s natural language processing would be a progression for mobile Web applications.

Imagine a user needs to know how to program his mobile phone to forward calls. He could access the frequently asked questions utility on the phone’s Web site and ask for instructions on the task. The speech recognition component would allow him to communicate with the Web’s virtual assistant,  accessing the user’s phone information to display or send the instructions to the user via text-to-speech. This interaction would take little time for the user, a live agent would not have been necessary, and the user could speak his need without using a small keyboard to type questions.

Creating a sound unified communications strategy and integrating consistent functionality across different channels can enrich a company’s portfolio of client amenities. Not only would it give the appearance of a unified solution, but it also could consolidate a company’s resources to lower costs and reduce customer churn. Integrating speech and techniques such as semantic parsing into different channels is what’s needed to take these customer interfaces to the next level and to distinguish a company’s image.


Aaron Fisher is director of speech services at West Interactive, where he oversees the design, development, and implementation of speech applications for the company. He can be reached at asfisher@west.com.


Page1 of 1