Chatbots are increasingly being used to cut costs and increase operational efficiencies in a variety of customer-facing industries such as retail, banking, telcos, and utilities. Capable of learning from their interactions with consumers, bots can now outperform humans in many areas of customer service.
There are obvious advantages associated with the use of chatbots but also many risks. While mishaps don’t seem as costly in the tech and retail industry, where the thirst for chatbot interaction is high, in the banking sector, a breach of trust could prove disastrous.
Companies must embrace chatbots if they want to meet the needs of future consumers but they should do so cautiously as chatbots are still, in many ways, a work in progress.
Millennials and the rise of the bot
Today, Millennials are playing a key role in the rise of chatbots with almost 60% saying they’ve used one and 53% of those who haven’t saying they were interested in doing so, according to a study by Retale.
The results also revealed that, among adopters, 39% qualified their experience as “very positive” and 31% as positive. Additionally, 71% of Millennials surveyed expressed “interest in trying chatbot experiences from consumer brands.”
This is hardly surprising since social media, where many of these experiences are found, is a pillar of millennial social life and has become a gateway for B2C interactions powered by chatbots, such as shopping and banking.
“Chatbots offer a sizable opportunity for brands to interact with this coveted audience,” explains Retale’s CTO, Dan Cripe. This “new form of one-on-one conversations between brands and consumers” embraces “a more conversational and personal approach to commerce and engagement that millennials seem to crave more than ever.”
However, there is room for improvement. Millennial respondents listed “accuracy in understanding what I am asking and looking for” as their biggest pet peeve (55%), followed by the “ability to hold a more ‘human’-sounding, natural conversation” (28%), and “getting a human customer rep involved where needed” (12%).
The challenge is to find the right mix of AI to human service, says the Retailer’s Guide to Chatbots, Live Chat, and Messaging report, because while consumers are increasingly showing “a willingness among consumers to adopt new channels such as messaging apps and chatbots, provided human interaction remains easily accessible when needed.”
When things go wrong
Despite their increasing popularity among consumers, chatbots still represent a risk. “Companies are racing to do chatbots and I think they’re risking real damage to brand equity,” warns Zor Gorelov, Kasisto’s CEO.
The problem is that chatbots can sometimes lead to costly mishaps. “A poorly designed chatbot can easily turn a potential customer engagement into a horrible user experience,” explains Mariya Yao, Head of Research & Design at TOPBOTS.
Microsoft’s Tay for example, which was created to respond to tweets and chats on social media, had to be turned off after only 16 hours. The AI, programmed to learn from its interactions, began to reuse misogynistic and xenophobic remarks made by users.
Human interaction still needed
In part, blunders such as Tay, have contributed to paint chatbots as inferior to human customer service teams and in many ways, they always will be. Human beings are more flexible and capable of making mental jumps that would seem illogical to a program.
As a consequence, Yao argues, chatbots should never be deployed “without establishing an escalation channel through which it can route customer issues it cannot adequately solve to humans who are trained to handle such issues.”
This is especially important in post-purchase service because consumers are used to turning “to other humans when in need of prompt assistance.”
The machine advantage
Despite the risks, the use of chatbots can have many benefits for companies when it comes to customer service. For example, they can:
Not only can bots accomplish many tasks more efficiently than humans, but they’re also becoming more and more personable, which can have both positive and negative consequences.
The “human” side of bots can facilitate transactions
Recent advancements in speech recognition will no doubt help improve the quality of chatbots but it’s unlikely they will ever be able to fully grasp human emotions in the way a person would. However, it doesn’t mean they can’t learn to mimic them through the use of sophisticated algorithms.
For example, Apple and Microsoft’s personal assistants Siri and Cortana have been “waging charm offensives, both quick to crack a joke or tell a story. Their elaborate personas are meant to keep users coming back.”
Many major tech players go to great lengths “to humanize their chatbots and establish a “real” connection with their audience,” says Yao. This makes sense since bots with loads of personality can make interactions more pleasant and fun. Who wouldn’t want to interact with Jarvis from The Avengers on a daily basis?
But “the visual elements (icon/avatar), vocabulary, tone, and overall personality” must resonate with the target group, warns Yao, because they represent the brand just as much as human employees do.
Beware the uncanny valley
On the negative side, bots imbued with personality can sometimes alienate customers and have been accused of misleading users into thinking they were digital “friends” rather than simply created to accomplish a task.
Some humanised bots can also trigger what’s called the uncanny valley. When robots are made to look more human, they trigger a sense of comfort in people. This same level of comfort exists when robots look exactly like humans.
However, “when robots appear almost exactly human” it leads to “an unsettling feeling that causes revulsion.” In turn, this can have consequences on the bot’s level of trustworthiness.
In a study conducted to assess the effect of the uncanny valley, participants were shown the pictures of 80 existing social robots and given $100 to distribute to the ones they felt would give them the best return on investment.
The results show a strong increase in the money given to robots, as they got more human looking, “before drastically dropping, only to increase again when robots began to look identical to humans,” thus highlighting the uncanny valley and its impact on the level of trust.
Trust, a key factor in chatbot deployment
The list of pros and cons of using AI-driven bots in customer service can make one wonder if they’re suited to every industry. Banks for example, pride themselves on their dependability and trustworthiness.
Considering trust has been found to be a consistent factor influencing customers to stay-purchase-recommend, according to Affinion’s The Connected Customer, isn’t there “a certain amount of risk to putting a chatbot out there that could make embarrassing or serious gaffes,” asks American Banker’s Penny Crossman.
The complexity of pleasing Millennials
Here again, Millennials can serve as a bellwether. While they’re viewed as the demographic most likely to embrace bots, their intentions actually vary a lot by sector. When asked about their comfort level with chatbots by industry,
The survey shows a clear correlation between the level of personal information needed and the willingness to use chatbots, with medical services – the most intrusive of interactions – at the very bottom.
It’s clear that chatbots are here to stay. There were 30,000 chatbots in operation on Facebook Messenger alone in September of 2016. The question is how can they be seamlessly integrated into customer engagement strategies and to what effect?
Integrating chatbots: a multi-step process
Chatbots today can accomplish a wide variety of tasks. Depending on the industry and its customer demographics, using a multi-step approach to bot deployment could be the wisest choice.
Consumers are a lot more used to interacting with automated systems in the telco industry for example, in part because they already use similar interfaces on their smartphones.
Building on that experience is a good place to start, explains Accenture. As customers get used to interacting with intelligent assistants, such as Siri or Alexa, “through a chat or voice interface, they will expect to do so in all their digital interactions.”
After that, through education and good design, customers can be led to more complex interactions yielding great advantages for companies.
Taking the plunge to foster engagement
With Aura, unveiled earlier year, O2 is taking its first steps into AI customer service. The new chatbot is meant to help the company cut its customer service costs by answering customers’ inquiries about their accounts.
In a few short months, the number of calls to its operator has dropped by half “as people increasingly turn to a smartphone app to manage their accounts.” O2’s goal, according to Chief Executive Mark Evans, is to continue to reduce costs but also to “increase loyalty by making it easier to interact with O2.”
Don’t forget to ask for feedback
According to Mariya Yao, Head of Research & Design at TOPBOTS, the most important aspect in a bot deployment strategy is to make sure to ask customers for feedback. Asking “an open-ended question requesting feedback after a key completed action enables developers to get valuable, qualitative insights,” she says.
Feedback can also be useful to increase sharing, according to Yao. When a customer or user selects a “Very likely” or “Likely” answer to a question, they should automatically get a suggestion to share with their friends.
Ultimately, asking “for feedback won’t always improve your bot, but not implementing intelligent feedback loops means your team will miss critical information used to design new features, fix frustrating bugs, and drive your user engagement,” she says.
In the retail sector, “conversational commerce” is described as the future of digital shopping. Already, “brands and retailers are able to connect with consumers in a personal way through messenger apps and chatbots.”
As they become even more sophisticated, chatbots will be able to scan a person’s messages, browsing history, and purchase behaviour to offer them a personalised list of products that’s sure to please.
The use and quality of chatbots will no doubt vary from one industry to the next, but one thing is for sure, those who handle personal information will have to work harder to convince consumers their bots can be trusted.
Ultimately, there’s no turning back the tide. Chatbots are here to stay and companies must start laying down the foundations of their AI-based customer service if they hope to meet the needs of Millennials, the largest group of consumers for many years to come.