To improve customer experience, it’s a matter of voice
From customer service to branded “personalities”, voice tech is expanding what it means to speak with target audiences.
Customers want to be heard, and the rise of voice assistants and audio channels are giving marketers more options to listen to them.
This trend has been gaining steam over the last year, as homebound customers reached out to brands through new channels they hadn’t previously tried. In 2020, nearly 30% of consumers called a brand by phone for the first time, for example.
Many of the voice and audio technologies have been around for years. But as brands have adopted them, either to manage high volume during the pandemic or to gain a competitive edge, these technologies might have earned a permanent home in marketing stacks.
“Technology kicks in to save you when the volume of customers rises,” said Chris Greenough, CMO of CX support technology company Everise. “During the pandemic, companies have scaled chat volumes.”
Responding in time
Warehouse retail chain BJ’s Wholesale Club began working with Everise in the months prior to the pandemic without knowing how much they would be depending on the technology to manage the volume of incoming calls.
The company’s platform uses natural language understanding to manage voice interactions with customers. The solution allowed BJ’s to understand what customers were asking for and either to provide easy-to-understand instructions or direct customers to a live representative.
The AI powering the platform has to be trained to understand context and regional nuances in language. It’s also being constantly re-evaluated and improved as use cases change.
“You do need both people and technology to run all of this,” said Greenough. “For conversational solutions, you still need people to train the data and give voice to it to improve how it’s architected.”
BJ’s stores are located mostly in the Northeast U.S. and down the East Coast, and the platform had to interpret different regionalisms and dialects within this relatively small area. It also has to be responsive to generational differences. For instance, an older customer might not immediately understand that “check your inbox” means that the brand had just sent you a support email.
“We launched the BJ’s project right when the pandemic hit the United States,” said Lana Pagnoux, Consultant for Everise. “The project was initially to streamline the digital experience operation. It was a bit serendipitous that the project wasn’t planned for COVID-19. The solution sort of answered any challenge they would have.”
Everise created the conversational solution for BJ’s along traditional KPIs, including resolving the best contact for an incoming customer call or message. “Once the pandemic hit, there was a whole different set of KPIs,” Pagnoux explained. BJ’s had new policies that were constantly changing around how customers could visit stores and what kind of pickup and delivery services they could provide.
Pagnoux and her team continued to analyze the conversational data that was coming in from an influx of calls, learning what the new customer service cases were all about. Customers didn’t just want to know if the nearest store had toilet paper in their inventory on a given day, they wanted to know if toilet paper was ever coming back, and when.
In May and June of 2020, BJ’s received almost 478,000 calls, of which over half were managed by a voice bot. Eleven percent of those were about the status of a digital order that were then transferred to a self-service flow. That way, customers could help themselves, and live representatives and voice bots could help customers with other requests. The platform shaved off over 30,000 requests in just those two months.
Better customer insights and reach
Not only does AI technology improve as it learns, but the company using it gets smarter about its customers, too.
“There’s a wealth of conversational data that we feel is really insightful,” said Pagnoux. Because Everise sits on the customer support side, they don’t see sales and conversions themselves. But brands can connect the dots and build data from conversations powered by natural language understanding.
For instance, Everise can connect with the omnichannel messaging solution Bright Pattern, Greenough said. This allows marketers to understand and operate voice and texts from WhatsApp, on social platforms, or through their own website and owned channels. For Everise’s healthcare customers, it links up with Zendesk, he added. From there, the data can be stored and managed in a CRM.
Messaging channels allow customers to contact brands for service requests. Brands can cast a wider net by tapping into committed audio channels.
“Where we’ve seen a lot of potential is where brands are enabling their community to say something, to talk,” said Ahmed Bouzid, Founder and CEO of voice startup Witlingo. (Previously, he was Head of Alexa’s Smart Home product at Amazon.) “Social audio has exploded. Voice is no longer within the confines of Alexa and Google Assistant.”
Voice engagement with consumers has piggybacked on the social space and through radio channel audiences, who stream content digitally, he said. Branded audio channels can be produced and controlled and are inherently less risky. However, live audio experiences like those on Clubhouse are also part of this frontier.
Podcasting is also going mainstream in a major way with easy production and monetization tools like Anchor, he added.
Voice search isn’t going away. In fact, it might finally gain more adoption as younger users are embracing it, said Bouzid.
The opportunity in assistants
“Amazon threw Amazon Echos out the door (for low prices) and got 100 million of those devices installed in homes,” said David Ciccarelli, Founder and CEO of Voices, a digital voice talent marketplace. “But brands didn’t catch up at all. Echos just became smart speakers for listening to music. The apps weren’t quite there yet.”
Ciccarelli applauds Crest for developing a storytime app, for children to listen to while they brush their teeth at bedtime as one example of how voice is maturing into a real marketing channel now that devices are in most homes.
For example, Amazon’s Polly solution allows brands to craft a unique voice by inputting phrases and building out from there. Brands don’t want to all sound like Alexa, after all.
“They’re basically creating another dimension of personality,” Ciccarelli said. “If you own the brand identity you can have different actors, as opposed to hiring a celebrity,” he added.
Platforms like Voices also allow marketers to discover and assign the right voice actors for multi-region campaigns.
Filling these many voice roles at scale can also help fuel creative optimization, a practice where AI pieces together the best elements of an ad, based on what a particular audience is most receptive to. (Although not available yet, audio is part of the roadmap for IBM Watson Advertising Accelerator, according to David Olesnevich, head of product at IBM Watson Advertising.)
“The cost of the ad is a fraction of what the ad buy is going to be,” said Ciccarelli. “So why not A/B test the concept? Might as well pay for another iteration of the ad and split test that in-market.”