Vertical Markets Spotlight: Speech in Retail

Article Featured Image

Speech adoption in retail continues to grow as merchandisers use the technology to provide consumers with another channel with which to buy products and services, improve call center interactions with customers, and enhance self-checkout, among other uses.

Much of the use of speech technology is incorporated into omnichannel strategies, with the technology being used to capture voice conversations, the data from which is combined with data from other channels to recommend products and services to customers and to help contact centers improve services for retail customers.

Conversational commerce, using voice for shopping, is the biggest speech technology development for retailers in the past several years. Voice assistants such as Amazon’s Alexa and Google Home are becoming more ubiquitous in the home, and the number of connected automobiles are growing as well.

UnivDatos Market Insights, a European research firm, in its latest research projected the global voice assistance commerce market to grow at a compound annual rate of 77.7 percent through 2027, when it is expected to reach $1.3 trillion.

Such widespread adoption, plus continued improvement in and use of speech technologies, has established good usage of speech technology for retail today and sets the stage for huge growth over the next several years.

“Operating in conjunction with other technologies, such as natural language processing, conversational systems have the potential to disrupt a variety of sectors by redefining the way businesses interact with their customers,” says Itonics, a digital innovation platform provider, in its “Game Changing Technologies for Retail” report. “Conversational systems of the future are likely to be more effective at self-learning and improving efficacy over time. Domain-specific conversational systems are likely to be developed to cater to specific industries. Similarly, solutions capable of capturing potential leads and making personal recommendations could transform sales processes.”

According to Capgemini, 87 percent of consumers with voice assistants have used them to order meals, and nearly as many (86 percent) have used them to purchase products. Half of consumers surveyed said they found shopping via voice assistants more convenient than making purchases through other channels.

More than four in 10 (41 percent) consumers would rather shop via voice assistant than a website or app because the voice assistant helps them automate their routine shopping tasks.

Retailers on digital channels can use conversational commerce combined with their own customer knowledge base to build compelling customer experiences, targeting customers who show a high affinity for voice and whose history has shown a higher affinity for buying through voice assistants, according to Capgemini.

Retailers that want to set themselves apart can use conversational AI to scale the personalized shopper experience, thereby differentiating themselves in the e-commerce market while driving brand loyalty, the Itonics report says. These virtual sales associates can be programmed to deliver highly specialized knowledge and experiential engagement, from serving as a technical product expert offering recommendations or step-by-step guidance to an exclusive concierge service.

One retailer that has already started such a strategy is skin care app/consultancy HelloAva, which uses a chatbot to ask customers a series of questions designed to guide their selection of skin care products. Based on the answers, customers will be put into one of 30 skin type categories from which the chatbot will recommend specific skin care products. The company has partnered with a variety of beauty brands to build its own inventory and checkout experience.

Some Consumers not Convinced

Despite the successes many retailers have had with the technology, the Itonics report has this word of caution: “This technology is still limited in that it struggles to interpret the precise emotion and tone of human input. Additionally, proponents of this technology need to address the public’s growing reluctance to share personal information with conversational systems.”

Capgemini agrees, saying that 74 percent of non-IVA users it surveyed were skeptical about using the technology for purchases, citing lack of trust that the technology would indeed place the order and accept payment correctly.

Nearly one-quarter (23 percent) said they were concerned that someone might impersonate them and order without their knowledge. Sixteen percent said they would like to try the technology a few times before trusting it to order on its own.

Despite this hesitancy from some, Capgemini says that active users of voice assistants expected 18 percent of their total expenditures to take place via a voice assistant in the next three years, a six-fold increase from today.

Additionally, large retailers like Amazon are relying on speech technology for targeted sentiment analysis, according to Clifton Wiser, vice president of solution architecture for TTEC’s Voice Foundry subsidiary. Unlike basic sentiment analysis, targeted sentiment breaks an experience down to its component parts. Rather than simply assigning a positive rating to a shopping experience, for example, the customer can rate the experience as above average, but rate the checkout time as poor.

“Targeted sentiment analysis really unlocks the true voice of the customer and the ability to do real conversation modeling at scale,” Wiser says. “In the retail space, that is gold because that VI data lake can inform ever-better customer experiences and ever-better customer products and services.”

Amazon introduced its targeted sentiment tool, Amazon Comprehend, in March. The natural language processing service uses machine learning to uncover insights in data from voice calls, chatbot interactions, emails, etc. The API provides more granular sentiment insights by identifying the sentiment (positive, negative, neutral, or mixed) toward entities.

The Amazon Transcribe speech engine takes the audio, selects the language model, identifies the active speaker, then breaks out an active transcript for Amazon Comprehend, which in turn determines targeted customer sentiment.

Amazon Comprehend isn’t exclusive to Amazon; rather it’s part of Amazon Web Services, which any retailer or other business can purchase. It can be integrated within a contact center’s ecosystem. Like other AWS services, users pay for Comprehend based on usage.

Capgemini recommends that retailers start small, taking a “test and learn” approach to driving customer use of conversational commerce. Further, they should focus on four areas to develop a robust conversational commerce strategy:

  1. Design and execute compelling voice experiences. Design the voice experience based on the target group’s needs and experiences. Align incentives to drive greater use of voice assistants and ensure that customers understand the value they can receive by using the voice channel for retail.
  2. Apply conversational intelligence to understand the enterprise’s target customers. Examine key customer journeys and assess how voice interfaces might improve the customer experience.
  3. Devise sound business solution operations incorporating voice solutions. Deploy a communication and marketing plan to build a loyal voice user base. Focus on CRM to ensure retention.
  4. Implement technology options that seamlessly integrate voice. Pull together customer data, technologies, and processes for the voice channel. Personalize the experience, but also take proactive steps to secure consumer data, especially for first-time users.

And speech’s use in retail moves beyond the initial sales channel, too.

“In retail more than any other industry, when people call, they have chosen voice as their channel and they want to talk to somebody,” says Rebecca Wetemann, CEO and principal of Valoir. Speech technology combined with AI can detect if the caller is speaking slow, fast, with a particular type of accent, etc., to route the person to an agent who can speak in the same voice.

The technology is moving from being an application to being a strategy in retail today, Wettemann adds. The data collected from voice interactions help contact centers with their training and planning for future interactions. “It’s about gathering better data, understanding that data, and having better interactions with customers for one transaction to several transactions.”

Speech Technology at Checkout

While much of the speech technology in retail is used for conversational commerce or in the contact center, the technology is also increasingly finding its way into more traditional brick-and mortar settings, says David Ciccarelli, founder and CEO of Voices.com, a marketplace for voice-over artists.

In Canada, for example, more retailers are turning to self-checkout lanes to serve customers better. According to Dalhousie University’s Grocery Experience National Survey Report, more than half of Canadians (54 percent) like self-checkout lanes, and 66 percent use them at least some of the time. “Self-checkout lanes allow customers to move at their own pace, but they’re not always as efficient. By incorporating voice technology, you can ensure that customers are guided through the checkout process quickly and effectively. Verbal instructions can help remind customers to swipe their rewards cards, weigh their produce, and pick up their change if they pay using cash,” Ciccarelli says.

Beyond voice-aided help at the checkout, more than a third of consumers would be willing to replace customer support or shop sales support with a personalized voice assistant to enhance their in-store experience, according to Capgemini.

In just those use cases alone, retail is a huge market for speech, and it is only expected to increase. 

Phillip Britt is a freelance writer based in the Chicago area. He can be reached at spenterprises1@comcast.net.

SpeechTek Covers
for qualified subscribers
Subscribe Now Current Issue Past Issues