We’ve all been there.
You’re on a company’s website when a chatbot pops up, chock full of AI to help answer your questions. Despite your best efforts to communicate, however, you end up confused and frustrated and request to speak to a real person.
It happened last year when Telstra customers flocked to social media to blast Codi, the Australian telecom provider’s new virtual assistant, as “stupid, ineffective and offensive” for not answering simple questions and ignoring their requests to be transferred to a human.
At other times, AI lacks a little … sensitivity. The UX Bear chatbot was set up to answer questions and discuss trends related to design. Here’s how one awkward exchange went:
UX BEAR: “How would you describe the term ‘bot’ to your grandma?”
USER: “My grandma is dead.”
UX BEAR: “Alright! Thank you for your feedback.” [Thumbs-up emoji.]
Ouch :- (
How can companies harness the business benefits of AI – like efficiency and cost savings – without completely stripping humanity from the customer experience?
The business case
Fortunately, we’re seeing fewer of these “chatbot fails” as AI algorithms improve. French sporting goods retailer Décathlon has had success with its chatbot and recently shared their story at DX3, Canada’s largest digital marketing and retail conference.
“It’s conversational commerce through a chatbot,” Alexandre Toulemonde, Décathlon’s chief technology officer, told the DX3 audience.
Montreal-based Heyday.ai developed a text-based chatbot for Décathlon, which opened its first store in Canada last spring. So far the tool has attained a 28 percent chatbot-to-sale conversion rate and a 92 percent customer satisfaction rate.
Despite these promising statistics, Heyday.ai CEO Steve Desjarlais posed this question to the DX3 crowd: “How can we humanize commerce in a mobile world with AI in it?”
For Décathlon, that means deploying chatbots to complement human staff rather than replacing them entirely. Toulemonde said chatbots “free up your (sales) team from repetitive work” – like checking if items are in stock or connecting shoppers with the right department – so they can focus on providing customers with a better experience in-store or over the phone.
Decathlon’s chatbot also displays a “talk to a person” icon on a customer’s device screen so they can immediately switch over to human help if they wish.
Some companies, however, are making their AI assistants so lifelike that it’s actually hard to tell them apart from humans.
Can you spot the bot?
Google’s new virtual voice assistant, Duplex, duped reporters into thinking it was a real person on the phone during a demo last May. A few months later, Leap Magic unveiled Mica, a visual assistant that interacts with users wearing augmented reality glasses. Mica’s appearance and gestures were so lifelike that testers began mirroring her behavior. They were smiling back at her and yawning after she yawned.
“Our goal is nothing short of the most realistic human experience in spatial computing,” John Moros, VP of human-centered AI at Magic Leap, told CNN.
Brands that want to break away from the pack should focus on emotion,Forrester, “The US Customer Experience Index, 2018“
ashow an experience makes customers feel has a bigger influence on their loyalty to a brand than effectiveness or ease in every industry.
The human edge
There’s a good reason companies are striving to make AI more human: customers respond to it. A 2018 Forrester survey of 110,000 U.S. consumers concluded that “brands that want to break away from the pack should focus on emotion, as how an experience makes customers feel has a bigger influence on their loyalty to a brand than effectiveness or ease in every industry.”
In the survey, brands rated highest by consumers provided an average of 22 emotionally positive experiences for each negative one; the lowest-rated brands provided just two emotionally positive interactions for each negative one.
Today’s consumers also want hyper-personalized experiences. To deliver this, companies are analyzing data about individual customers so they can treat each one like a longtime friend, not just part of a demographic.
“Ninety-one percent of consumers are more likely to shop with brands who recognize, remember and provide relevant offers and recommendations,” according to Accenture’s 2018 global survey on digital personalization.
Getting too personal
Yet brands can also get too close for comfort. In the same Accenture study, 27 percent of consumers cited “a brand experience that was too personal or invasive.” One in four said it’s “creepy” to receive a text or mobile notification from a brand as they walk past a physical store.
As technology becomes more intelligent and lifelike, questions arise about privacy and ethics. Under the law, companies have to notify us and obtain our consent before collecting and using our data. Should they also be required to inform us when we’re talking to a bot instead of a real being?
For now, it’s a question that humans can still tackle better than Siri and Alexa.