Rise of the Voice Assistants

  • March 04, 2019
NTT DATA Services Virtual Assistants Blog

Voice assistants are creating a buzz reminiscent of the rise of smartphones a decade ago. This emerging technology can be built into a variety of devices and can be used in nearly any environment to expose functionality in new ways and types of interaction. Platforms, such as Amazon’s Alexa and the Google Assistant, enable companies to provide solutions that can help customers and employees achieve their goals quickly and easily.

Research conducted by NTT DATA Services found that there is a growing constituent of consumers who prefer to interact with businesses through voice assistants. As such, voice assistants represent an important segment when considering automation solutions. So, how can your business best prepare to accommodate this growing population? And what should you consider when making the decision about the scope of voice assistants within your business model?

NTT DATA recently completed an Alexa skill project for a leading independent mortgage lender. From this initiative, we have compiled five best practices that can help businesses when implementing voice assistant solutions. Here are two recommendations:

Be concise

According to the study, consumers are more likely to interact with a voice assistant when the interaction is quick and simple. So organizations should target use cases that provide information to help users make decisions quickly. Users should be able to speak their requests intuitively, and responses should avoid both verbose feedback and back-and-forth interactions to obtain additional details. This need for quick, succinct interactions may result in requests being broken up into smaller chunks. Information that may be grouped together within a single interface on the web or a mobile app may need to be served up through discrete interactions within a voice assistant so that just one particular data element is obtained at a time.

The same study found that consumers still value the accuracy of a response over speed and convenience. NTT DATA’s recognition of — and support for — these consumer preferences is evident in the solution we recently developed, which supports concise use cases that provide details about loan applications. We structured the skill to respond accurately with the desired information in only one statement. Additionally, we used features of the Amazon Alexa Dialog Model to obtain the criteria necessary to avoid further prompting. When the result set is reduced to no more than three items, those items are listed for the user to choose from. This model not only presents the best opportunity to fulfill the request entirely, but also provides a simple interaction pattern for obtaining additional information, and limits the amount of response required when the user must select from a list.

Adopt a flexible approach

One of the key features of voice assistants is their support of natural language processing, which means they need to be able to accept requests spoken in the variety of ways humans naturally speak. Some people may want to interact with voice assistants by providing all the required details of the request in one shot, whereas others may prefer to have their assistant prompt them for the required details. Users can even initiate the dialog in various ways. For example, they can ask to do something directly within the initial request, or they can “open” the skill first and then specify what they want when prompted.

For direct invocation requests, additional details can be incorporated to support a more natural interaction. The mortgage skill NTT DATA developed highlights this capability, where the skill enables the user to make a request such as, “Alexa, ask My Mortgage Skill for the status of the Johnson application.” In this case, the invocation provides the information necessary to completely fulfill the request (skill name, type of request and name associated with the request). Of course, the user could have conducted the same interaction as a multistep process, so building the solution to accommodate various types of dialogue allows the skill to support use cases in the most natural way possible for the user.

Looking toward the future

The rise in popularity of voice assistants affords new opportunities to interact with customers and employees. However, you need to be aware of the expectations your users have within those scenarios and deploy solutions that provide them with intuitive and seamless experiences. Incorporating the best practices described in this white paper will help you create solutions that support these goals.

Read the five recommended best practices in What to Consider When Considering Voice Assistants.

Subscribe to our blog

John Fleming, NTT DATA Services
John Fleming

John directs technology innovation at NTT DATA’s Collaboration Center in Plano, TX. He has more than two decades of experience in software solution architecture and implementation, with a primary focus on mobile, web and IoT. His responsibilities include identifying and managing innovations within the Collaboration Center and supporting client engagements involving emerging technologies. John has also spearheaded multiple R&D initiatives at NTT DATA, helping drive the direction of IoT, wearable computing, mobile and web capabilities within the organization.

Related Blog Posts