When interviewing potential new employees, you appraise them on multiple levels. Like how they present themselves, what they wear, how they speak and what they say. You ask yourself, “Can I work with this person?” Or more importantly, “Do I trust them representing our brand in customer-facing scenarios?" These same questions and concerns should apply when selecting and branding your personal virtual assistants, or chatbots, experts told CMSWire.
According to Bynder’s 2018 State of Branding Report, 40 percent of marketing decision-makers said they planned to invest in voice assistants this year. And, naturally, they care greatly about how that virtual assistant will present itself to customers. It’s your brand that’s on the line, just like with your human employees.
The data showed that the tone of a voice assistant ranked more importantly than gender when it came to branding faceless voice-enabled technologies, although 54 percent of marketers would still prefer a female voice assistant.
What Is a Virtual Assistant?
To be clear, we’re not talking about a human assistant that works from home. This is an Artificial Intelligence (AI)-driven virtual assistant that represents your brands in a digital experience and often is on the front lines of customer support, service and experience. These online helpers are commonly referred to as chatbots. “Chatbots and virtual assistants (VAs) may be built on artificial intelligence and deliver digital experiences through virtual personas — but the success B2B marketers realize from them will depend on the real and human aspects of their deployment, intraorganizational impact and customer obsession,” Forrester researchers wrote in a report on chatbots this year (fee charged).
With so much on the line, what should you consider before rolling out these digital helpers? We asked some experts for tips on branding considerationsthat will help ensure a chatbot/virtual assistant roll out is successful.
Related Article: 6 Success Factors for Workplace Chatbots
Set Realistic Expectations
Joe Hearty, associate design director at W12 Studios, said it’s important for brands to set realistic expectations by clearly defining and communicating what it is your chatbot or voice assistant actually does. Establish a primary function or target user group, and, from there, identify which use cases feel relevant and start designing with them in mind, Hearty said.
Make Your Chatbot Consistent With Brand
The very first question companies should ask when it comes to using a virtual assistant is: “Am I offering a consistent brand experience across the channels my company serves?” Jane Price, senior vice president of marketing for Interactions, shared that thought with CMSWire. “Effective virtual assistants help companies deliver productive, positive and personal customer experiences,” Price said. “However, many companies find this a struggle, because keeping up with emerging channels and consumer demand has forced companies to cobble together various solutions, losing consistency and even brand voice in the process.”
Make sure your virtual assistant is consistent with the overall brand experience your customers have come to know and love. Ensure the overall brand experience is translated across all channels — phone, voice assistants, webchat, SMS, social media and more, according to Price.
Related Article: 8 Ways to Measure Chatbot Program Success
Your Chatbot Persona Should Represent Your Demographic
When branding an assistant, you’re creating the personality of your most visible customer service representative, said Robert Weideman, executive vice president and general manager of the Nuance Enterprise Division. Create a chatbot that embodies your brand. “For example, a bank positioned as ‘secure and reliable’ may want to craft a persona that is more serious and evokes being trustworthy and dependable, whereas a large retail chain that targets Millennials may choose a personality that’s more relatable, humorous and engaging.”
Brands investing in virtual assistants must consider the delicate politics of gender, personality and accents when developing a (literal) brand voice, and “be sensitive to any underlying bias,” said Nicolas Deskos, head of content at Bynder’s OnBrand. "There’s a lot to be mindful of when branding a virtual assistant — including tone, speed, accents and personas, catch phrases and slang, the age of the assistant and even the gender.”
Represent the Diversity of Your Customers
The team at Sage have a different approach to their chatbot persona. They have developed an AI-powered smart assistant called Pegg that is gender-neutral. “[Pegg is] proud to be a robot and does not need a gender to manage the business management tasks asked of it,” said Kriti Sharma, vice president of artificial intelligence at Sage.
Sharma added that gender bias in AI reflects the challenges we’ve faced as an industry for a very long time. R&D teams, she said, are just not diverse enough, with fewer women having the opportunity to participate and ultimately influence tech innovation. “Technology needs to represent the diversity of its users, and we will only truly reach this utopia when the teams that build both business and consumer technology are representative of society as a whole,” Sharma said. Brands that continue to flag these industry-wide unconscious biases can help stop the perpetuation of these “tropes in the technology we create."
“[At Sage] We are working to make the AI technology we consume at work and at home more gender diverse, than what humans have made for themselves,” she added.
Ensure the Data is Available to Deliver on Expectations
Remember, the color choice for your virtual assistant’s clothes and tone of voice may be important, but it needs to deliver service that is fast, accurate and more convenient than alternatives, Weideman said. “Out of the gate,” he added, “an AI assistant needs to have a foundational understanding of user intent, and be designed with pre-loaded, industry- and business-specific knowledge and expertise. It must be able to leverage past experiences to deliver answers with a wide range of knowledge. It also must be able to drive a high confidence level in its ability to assist with more complex and domain-specific activities.”
Understand All Conversational Tones, Engagement Channels
AI assistants fall short in conversation and dialogue via text and speech engagement, Weideman said. Natural Language Understanding (NLU) and conversational dialogue systems are not created equal, he added. “The best are able to really understand what the user says or types (intent), ask clarifying questions (dialogue) and remember the context of the conversation,” Weideman said. “In short, they mimic human conversational intelligence. This allows consumers the ability to engage with self-service as easily as with contact center agents, even when the contact center is closed.”
Chatbots should also have what Weideman calls an “intelligent engagement brain” across all channels that consumers use to interact with companies — Web, mobile app, text/SMS, phone, messaging apps, smart speakers, TV, car and more. “In doing that,” he said, “the enterprise will see higher levels of customer satisfaction and increased revenue, all with lower costs of engaging with their customers.”
Related Article: 8 Brands Innovating with Chatbots
Ensure Emotional Relevance
According to Tara Kelly, CEO of SPLICE Software, organizations that brand virtual assistants should ensure emotional relevance. “When reporting a break-in to your insurance company,” Kelly said, “the last thing you want to hear is a chipper, upbeat robot voice because it feels as if the company isn’t taking your situation seriously. Using relevant, emotional tones in the voice of virtual assistants is crucial if you want your brand to forge long-term connections with customers.”
Consider Designers, Constant Testing
Sound is an element of voice interaction that is often overlooked, so if possible, have a sound designer involved from the get-go — along with healthy spread of interaction designers, visual designers and front-end developers, Hearty said. “Not only will this result in a richer experience for your users, it’ll also help in fine-tuning more granular elements of your experience such as audio fidelity, timing and tone of voice,” he said.
Further, rapid prototyping and testing is “absolutely crucial for stress testing the logic that governs your experience,” Hearty said. “Language is a complex beast, not only will you have to accommodate the expected inputs for your ideal use cases, you’ll also need to account for a myriad of edge cases produced as a result of mispronunciation, poor grammar or variations in accent or speed.”