Woman using voice assistant.
Sep 01, 2022
5 MIN READ

Don’t Give Away Vital Customer Feedback—Brand-Owned Voice Assistants Give You Access

If there’s one thing we’ve learned about what consumers expect from their devices and apps, it’s this: they want it convenient, simple and intuitive so they can do things fast. And don’t expect them to tolerate clumsy UIs. The most successful brands today from Apple mobile devices to simple-to-use banking apps to HR apps like Bamboo, have learned this lesson well.

A similar dynamic is now driving voice. Consumers are responding positively to voice because, when done well, voice is giving them more convenience and simpler functionality. And it’s fun. It’s why many brands across industries are looking for ways to add voice AI to their products, services, and apps. 

A study by Adobe revealed that 91% of business decision-makers are significantly investing in voice AI technologies, and 94% of business leaders plan to increase their investment in voice. 

How Voice AI Meets High Consumer Expectations

The question for brands, though, as you look to add voice to your apps, products and services, is how can you meet or exceed your customers’ desires. 

One clear differentiator is the ability to have direct access to your customer data. How else can you understand your customer’s experience—if they are happy, frustrated, delighted, or disappointed? Where do breakdowns occur that discourage users if not remedied quickly? 

Ultimately, do you have access to data so that you can respond quickly to adjust your  voice assistant to nuances in your user’s speech, update your content library to respond to requests that were not recognized, and keep users delighted with the experience. 

Which Voice AI Strategy is Best?

We’ve identified 3 pathways to adding voice AI to your apps and devices: you can build it yourself, “rent” it from a big-tech provider, or partner with an independent voice AI provider.

While each can get you there, if a key focus is your customer’s experience, consider this example and which option best allows you to do this:

Say you are a music subscription company and you’ve designed your branded voice assistant to respond to user requests by the song’s title. But you start to notice a drop in the number of songs being played. With direct access to the transcripts and queries of customers, you learn that many users are not, in fact, using titles to request songs, but lyrics—a capability not included in the voice assistant’s understanding. The impact may not only be less usage of your service, but customer frustration that can lead to higher unsubscribes.

With the knowledge and direct access to user data, your voice AI experts can modify the script to respond to requests by lyrics as well as titles.

“With a well-engineered voice assistant, you can look at areas of data to identify any problems, from the core speech recognition (ASR) to the natural language understanding to the audio path.”

Darin Clark, Director of BD at SoundHound

This type of trouble-shooting and diagnostics would be far more difficult to rectify if you rented a voice assistant from a big tech provider because access to data is too limited and the solution would be harder to implement. 

Retain access to customer and market data

When you sign-up with an established, big tech voice assistant provider, you’ll have very little access to important customer and market data. This is because most big-tech providers don’t provide the same level of data transparency and insights as an independent platform—and often may use it for their own understanding of market and consumer desires. In fact, it’s not unknown that some big tech providers have used the data to optimize their own products at the expense of their own customers. 

While losing visibility into your customer interactions may seem a trivial price to pay for a voice AI technology that doesn’t require much investment, the data behind your voice assistant could very well be your ticket to continued competitiveness in the market.

Brands looking to voice-enable their products must ask themselves one important question:  Do we want other parties to have more visibility into our user behaviors, desires, and preferences than we do—giving them the ability to leverage and benefit from our user insights?

You want to know how customers are using your voice assistant,  what questions they are asking, and what functions they are trying to perform. Data that is collected and anonymized can be used to inform not only the iterations necessary to improve your user experience but also your short-term and long-term product and services roadmaps.

“Data is one of the most valuable assets to a company today, and it will be only more so in the future,” says  John Goscha, Founder & CEO of Native Voice. “With our customers, they want a direct connection to their customer because they want that data. Ultimately, that’s how they know their customer and how they can customize their services.”

What customer data tells you

When your brand has visibility into data insights—including the most common queries, queries that are supported but need custom commands, and those that aren’t supported yet—you can use these data points to inform improvements in the voice experience. 

When the voice interface is used to operate a device, for example, manufacturers can easily see what additional functions users would like the product to have and make improvements in subsequent models that further differentiate their products from those of their competitors. 

Voice AI data insights

A good analytics capability from a brand-owned voice assistant should provide you these types of parameters:

  • Rate of accuracy
  • Speed of delivery
  • Frequency of false positives or negatives
  • Responsiveness in noisy environments
  • Ability to understand the accents of your target audience
  • Understanding of vocabulary words
  • Types of user queries
  • Frequency of queries
  • Use of domains
  • Product functionality requests

General data points, such as accuracy, speed, and false positives or negatives, will give an overview of how your voice assistant is performing. By delving into noisy environments, accents, and vocabulary words, you can look closer at how your audience operates and fine-tune the experience for your users. For example, your target audience may use particular words and they should be incorporated into the voice assistant’s lexicon. 

Looking at domain usage, type and frequency of queries, you will get an understanding of what your users are asking of your voice assistant and if they are finding it helpful. 

Public domains, for example, often include widely available data such as local search, recipes, traffic, sports, stocks, flight travel, and more. Specialty domains, on the other hand, include information only users of that product, service or app might need, such as for an industry. Perhaps a hospitality domain will include room service menu, activities at a resort or on a cruise ship, terms like canopy ride, flora and fauna, exotic fish names.  

The richer the voice assistant’s vocabulary, the more free flowing and satisfying the user’s experience will be.

Product functionality requests are also an important data point to examine to find out how your customers are using your product and what they are asking it to do that it doesn’t yet do. This information will lead to crucial product improvements that can inform the product roadmap in much the same way as expensive market research can.

Like any product introduced to the market, voice assistants are never one and done. They require monitoring and iterations based on user experiences to continuously improve. If your users are constantly frustrated by false positives, lack of accurate responses, or unresponsiveness to certain questions, they will likely abandon your voice assistant and possibly your brand.

Ability to express your brand in the voice assistant

Keeping your customer at the center of your voice design is critical to implementing a voice interface able to respond appropriately based on the user’s situation, emotional state and desired outcomes. The ultimate goal is to create a voice assistant that’s personalized, authentic, human and relatable.

When finding your brand’s voice, consider these elements: 

  • Developing a voice persona
  • Incorporating personalization
  • Owning your voice assistant
  • Getting stakeholder buy-in

Data retention gives you a window into the success of your voice assistant, the frequency of certain queries, and the times when your users are becoming frustrated. The adage “Knowledge is power,” is true for brands that own their voice-enabled device or app’s user data. Your customers will appreciate the continued iterations of your voice assistant that make their experiences better over time, solving for the challenges they encounter, and answering the questions and providing the functions most important to them.

The very nature of voice assistants allows humans to connect on a more personal level with the world around them—by talking to devices and brands as they would other people. Personalized experiences are a natural extension of these relationships. 

Future: sentiment analysis

While Voice AI has seen numerous technological advancements, it is still young and in a highly innovative phase. New technologies on the horizon include the use of specific keywords and sentiment analysis that enable the voice assistant to recognize user moods and adjust accordingly.

Rather than respond by repeating the same request when the voice assistant doesn’t understand a user’s request,  it will recognize distress in the user’s language through specific keywords (“help,” “I’m finished,” or more colorful terms) and alter its approach, perhaps asking the question in a different way or altering its length of response or tone.

With sentiment analysis, the voice assistant can sense the user’s tone and respond in a more human-like way: “I’m sorry, I’m new at this” and reframe the question in a different way. 

The key point is that the ability to create successful engagements with your customers and strengthen your brand will depend on your awareness of what’s working and what’s not with your voice assistants. The more access you have to this data, and the more technologies at your disposal, will ensure customers view your brand in the best light.

Speak with an expert Request a demo

David Barry Headshot
David Barry is a high-tech storyteller with a passion for communicating complex ideas simply. He met his wife in a taxicab in San Francisco.

Interested in Learning More?

Subscribe today to stay informed and get regular updates from SoundHound Inc.