Chatbots: The pros and cons of automated customer service
“How may I help you today?”
Ten years ago, that message might have come from a human customer service agent on the other end of a phone line. Today, it could appear in a pop-up chat box as a consumer browses a website — and increasingly, that message might originate with a programmable chatbot.
Considered artificial intelligence (AI), chatbots are software that can be programmed to respond to people’s queries. TechWorld characterizes the chatbot as “a user interface which can be plugged into a number of data sources via APIs so it can deliver information or services on demand, such as weather forecasts or breaking news.”
Chatbots, however, have a distinct advantage: They let brands be — or at least appear to be — more proactive in addressing potential customer questions and concerns by acting as an integrated part of the shopping/browsing experience. The chatbot is right there, waiting to answer your questions, and then to connect you to a real person for more complex queries if necessary. There’s no 1-800 number to dial, and there’s no waiting on hold.
This is the promise of chatbots and the reason why companies are investing heavily in the technology. According to a recent MarketsandMarkets report, the smart advisors/chatbots market was estimated to be worth $703.3 million in 2016 and is projected to reach $3.2 billion by 2021 thanks to “the strong need to understand consumer behavior, adoption of cloud-based technology and proliferating demand of intelligent customer engagement.”
Learning valuable chatbot lessons from Clippy
If chatbots feel familiar, that could be because of the pioneering paperclip known as Clippy, the pop-up users used to encounter when drafting a document in Microsoft Word. The frequency of the software interruptions, and the questionable helpfulness, turned Clippy into a symbol of derision. Still, Clippy lived 10 long years, from 1997 to 2007, before the software company decided to lay him to rest.
Even if his constant interruptions were annoying, Clippy did teach us a couple of valuable lessons — namely, that chatbots are there to serve customers, not the other way around. And, as Clippy demonstrated, the line between useful and disruptive, or supportive and stiff, can be difficult to straddle. Mastering that delicate balance, though, is vital to the success of chatbots.
The pros of chatbots as contact center assistants
For Gam Dias, a principal at bot-design firm 1080Bots.com, having chatbots answer customer questions can help ensure a satisfactory customer experience, while adding a “sizzle” that shows the brand is up on the latest consumer tech trends. As Dias notes, chatbots are a great way to clear easy-to-answer questions off the plates of customer service agents, so that they may focus on more complex queries.
For example, Dias points out that around the holiday period, upwards of 70 percent of inbound contact center customer service calls are simply order status inquiries. This is where chatbots can excel.
A chatbot could take an order number, cross-reference that order with delivery status and provide the customer with an answer within seconds — faster than any agent on the phone could act. “If an AI agent can help 25 percent of those customers, that’s a huge savings for the retailer,” Dias says. “For those customers who are helped, it means they haven’t spent 15 minutes waiting on the phone for a status update.”
Chatbots are only as good as they’re programmed to be
Despite chatbots’ promise, not everyone has jumped onto the bandwagon. Scott Sachs, customer service consultant and president of SJS Solutions, likens the chatbot revolution to the advent of automated voice systems back in the day. “Everyone thought [voice systems] would solve everything and they wouldn’t need people,” he says.
In fact, he continues, automated voice systems — and chatbots, by extension — have generated an unintended consequence: “The easier issues are cherry-picked out, creating a more challenging environment within the customer-service organization, so they’re only getting the harder questions.” That means a higher skill set is required for frontline agents.
Additionally, chatbots aren’t all-knowing. They’re only as good as the programmers make them, and they can’t pivot as easily as a person can. Call a customer service desk, ask for a coat in the color lilac and an employee will likely direct you to a purple jacket. Tell a chatbot you want something lilac, and unless the company specifically programmed that bot to equate “lilac” with a light-purple product, the chatbot could come up empty — even if a light-purple coat exists. Your business has lost the sale, and the customer feels frustrated.
Some snafus in the past couple of years have also given chatbots a bad rap. In 2016, Microsoft’s Tay bot was swiftly baited by Twitter users into spewing hateful commentary. “It took less than 24 hours for Twitter to corrupt an innocent AI chatbot,” wrote the Verge. While the Tay bot may have produced some unintentionally offensive and even comical results, the take-home lesson here is that chatbots can be easily confused by certain questions and commands — and that could be more trouble for a brand than it’s worth.
Despite these growing pains, chatbots aren’t likely to go away anytime soon. One of the main reasons for chatbots’ presumed longevity is that they are bred to respond to online queries. As e-commerce grows, so too do the demands for customer service. At the same time, the artificial intelligence, machine learning and natural-language processing that power chatbots are growing by leaps and bounds right now. This means chatbots will only become more sophisticated and adept in the future.
How chatbots are helping brands right now
Although chatbot technology has yet to be perfected, brands are still taking this moment in time to experiment with, and optimize the technology.
Airbnb has utilized Amazon’s voice assistant Alexa as a tool to help welcome guests and teach them about the area. Third-party developers, meanwhile, have been working on easing communication between hosts and renters by developing chatbots that can respond to basic vacation queries on behalf of the hosts, according to VentureBeat. The bots automate answers to commonly asked questions such as, “does this listing have a dishwasher?” and “where is the closest subway station?” These are easy enough to program, and can save hosts time.
Elsewhere, makeup brand Estée Lauder has a chatbot within Facebook Messenger that used facial-recognition software to help users pick the right shade of foundation. The service has drawbacks, though: A simple query matching skin tone to available products seemed to work, but the chatbot couldn’t answer any additional questions.
The beauty of bots, though, is their ability to merge automated AI help with real, live humans. For instance, a customer service agent could take over from a bot if questions get too complex, or if the client has a high-value shopping cart in the works. This can happen seamlessly without further action from the customer. It’s in moments like these where bots show their true potential — not as replacements for humans, but as great assistants.