Adoption of smart speakers and other devices with built-in virtual assistants will create a big demand for customer support that’s simply spoken out loud.
OK Google. Ask [Retailer] where my order is.
Alexa. Ask [Airline] if my flight is on time.
Smart virtual assistants are literally everywhere – our phones, stand-alone devices sitting on our countertops, TVs, refrigerators, cars. We no longer have to lift a finger to learn the weather, understand the traffic on our commute or convert a measurement. And soon, getting customer support will be a few spoken words away.
Companies like Domino’s jumped in early to enable various parts of the customer journey hands-free. People can order a pizza from Domino’s without lifting a finger and clicking through an app, navigating the web or speaking with a human on the phone. WestJet now offers guests a new hands and screen-free way to get answers to questions about flight status, travel documentation and baggage allowances. It’s effortless on overdrive.
More and more, companies are starting to enable various tasks through these incredibly popular voice channels. The most exciting opportunity that solves a real business problem, though, is providing customer support on devices like Alexa and Google Home.
In this post, we’re going to explore the support opportunity on voice, ideal use cases for the unique channel, how to create an engaging experience, and considerations around privacy and brand safety.
One in four Americans own smart speakers like Alexa or Google Home1. People don’t just own smart speakers, but they are using them frequently: 27% use it nearly every day and 33% report using it multiple times a day2.
The incredible adoption rate has been paired with a change in consumer preferences for customer support. In just a few short years, 34% of consumers now say they would prefer to receive customer support via a smart speaker as opposed to calling a customer service line to speak to a human3. When you consider that smart speakers are still in their infancy, this rate of change is significant and signals a real opportunity for companies to resolve issues and answer questions via this channel.
When we start to think about the types of situations that could be resolved on voice, it’s important to understand how people are using the devices and what they are comfortable doing on them.
Three out of every four consumers (74%) are using their mobile voice assistants at home and primarily for basic tasks3. These include checking weather or news, setting an alarm or reminder, or looking up something that was previously done on a search engine. We’re not turning to voice assistants in overly complex situations. In fact, The New York Times found that people are often multitasking when they interact with their smart speakers: “People are using smart speakers most at transitional moments, like getting ready for the day and preparing dinner once they get home1.“
With all of this in mind, companies should identify high-volume and simple support queries. This aligns perfectly with the use cases that should be delegated to an AI Agent. Here are a few topics that would work extremely well on voice:
- FAQs: 59% of people have asked a quick question to a smart speaker3. Companies should identify the types of FAQs that people ask the most, such as baggage policies for an airline, check-out / check-in times for a hotel, or refund and return policies for a retailer. Don’t overcomplicate with too much information, but provide quick, clear answers.
- Status: Providing real-time status of a reservation, flight or order also represents great use cases for voice. This may require an additional level of authentication to provide the individual’s exact information, but is also a typically high-volume use case that would provide convenience for a big percentage of customers.
- Modifications, Cancellations and Upgrades: Allow customers to make meaningful changes to an existing order or reservation. For instance, an airline can allow a person to ask if there are upgrades on their flight this afternoon while they are zipping up their suitcase. Similarly, a subscription box company could allow a person to skip a delivery without having to log-in to their account on the web.
If you build it, they may come, and they may or may not come back.
There’s a lot more than figuring out the ideal use cases for voice and training an AI to respond. You also need to consider the personality, promotion and safeguards to keep customers happy.
The most critical thing for a voice assistant is its ability to properly classify a person’s intent. In a survey, over 50% of respondents say “How well it understands me…” is the most important factor in determining the user experience4. This accuracy will evolve during training, ideally from historical support data from other channels, but perhaps most importantly, call center data.
That being said, you also need to consider the personality and conversational elements of the virtual assistant. Is it funny or serious? Is it chatty or direct?
In a study, Mirroring to Build Trust in Digital Assistants, researchers found that 70% of participants preferred interacting with a chatty digital assistant as compared to a non-chatty one5. However, “the likability and trustworthiness of a digital assistant improves when the assistant mirrors the degree of chattiness of the user.” Training an AI to “mirror” the personality of a user can be done based on things like mannerisms, number of words use, frequency of interaction and tone. Eventually, companies can have virtual assistants that can take on a few different personalities based on the user – answering a question more direct or more chatty, with humor or without, and provide an experience that the customer loves.
As with an AI agent on any channel, you must provide a clear way to connect with a human if the customer grows frustrated or the AI can’t understand what the user needs. Train the voice assistant to read out a phone number or offer to send a text or email if it senses a user is irritable or gets stuck.
Companies will also need to educate people that support is available via these voice platforms. Having the information on a Contact Us page is a very basic example, but critical because discoverability is solely based on a person explicitly enabling or opening a company’s skill.
First of all, there are often multiple users of a single device, including over 2.2 million children who are playing music, asking questions and hearing jokes6. Research shows that children are not misusing these devices, but brands need to ensure all content is suitable for younger audiences. Additionally, put the proper safeguards like double authentication in place to ensure children are not empowered to make changes on an account, spend money, etc. without parental consent.
Additionally, people are weary about surveillance and spying. According to Pew Research, “54% of smart speaker owners say they are very or somewhat concerned about the amount of personal data their speakers collect3.” Ensure that when your skill is enabled, you are not collecting or storing any unnecessary data.
Voice is a part of a larger omni-channel customer support strategy
Voice is an exciting opportunity: it’s the newest channel that certainly has the “cool” factor. With customer demand for convenient support on the rise, it’s compelling for many companies to offer answers to questions and resolve common issues without a customer having to lift a finger. That being said, we don’t recommend launching a Support AI on voice. For historical and ongoing training purposes, ROI and business value, text-based channels like email should be where companies should get started with AI. Once an AI has learned and improved over time, voice is a great channel to scale to.
Do you want to try AI before you buy? We’re now offering a free 30 day trial. Start yours today.
- The New York Times: https://open.nytimes.com/how-might-the-new-york-times-sound-on-smart-speakers-3b59a6a78ae3
- National Public Media: https://www.nationalpublicmedia.com/uploads/2020/01/The-Smart-Audio-Report-Winter-2019.pdf
- PWC: https://www.pwc.com/us/en/services/consulting/library/consumer-intelligence-series/voice-assistants.html
- Voicebot.AI: https://voicebot.ai/2018/12/13/consumers-want-understanding-over-personality-from-voice-assistants-games-arent-on-their-radar-new-data/
- Arxiv.org: https://arxiv.org/pdf/1904.01664.pdf
- eMarketer: https://www.emarketer.com/content/the-smart-speaker-series-kids-teens-infographic