A chatbot with a megaphone, shouting in a customers ear
PHOTO: Shutterstock

All Aaron Dun wanted was to surf a website. There was only one problem: a chabot wouldn’t let him. It kept popping up during his web session, despite Dun’s multiple attempts to “X out" the chabot. Dun, a senior marketing executive in the Boston area, chronicled his web-travel chatbot woes in a popular LinkedIn post last month. “Please, please, PLEASE help your customers know when I have hit the little "x" to close the chat window, and you know, maybe help them WAIT awhile for me to decide I need help?!!” Dun wrote.

Dun’s LinkedIn post brings to light that chatbots can be powerful customer service agents. Yet, with all the hype and promise of customer-service chatbots, organizations that deploy them must build them logically and recognize they’re dealing with humans. And for most humans, popping up every few seconds when all you want to do is browse — think the retail person who follows you into every clothes aisle and asks how they can help — is not a great experience.

“Your systems need to be smart enough to know that I just clicked the 'X' box,” Dun said in an interview with CMSWire. “When I am in the same web session, and go to Page 2, 3, 4 and 5 in that session I don't want that bot popping up every time I go to a page. It should be smart enough to know that I closed it once.”

Related Article: Why Your Approach to Chatbot and IVR Projects Is All Wrong

It’s an AI, Chatbot World

One thing’s for sure: companies will have to get adjusted — and fast — working with chatbots and other forms of artificial intelligence (AI)-based customer service interactions. There’s no denying that chatbots remain super popular in the customer service arena. They’re also invading marketing, sales and HR, according to Julian Harris, head of technology research for AI startup CognitionX, who wrote about the chatbot topic in a Medium post

Some even see organizations moving beyond pure chatbot deployments and leveraging a full conversational AI program for customer-service programs. According to Amazon developer officials, Conversational AI systems “are computers that people can interact with simply by having a conversation, our most natural form of interaction. In short, it is what allows us to talk to voice-driven technologies like Amazon Alexa and ask about the weather, order products online, and even call a cab, simply by using the language we already know.” 

Rob High, chief technical officer at IBM, wrote this week that conversational AI can interact with customers directly to solve their most frequent problems and questions. "And, unlike many of the chatbot experiences on the market today, a conversational agent can seek to get to the question-behind-the-question — your client’s actual problem, not just the first thing that they say to open the conversation.” 

But chatbots aren’t going anywhere, according to industry numbers. Gartner earlier this year said 25 percent of customer service and support operations will integrate virtual customer assistant (VCA) or chatbot technology across engagement channels by 2020, up from less than two percent in 2017.

Related Article: 3 Considerations When Training a Chatbot

Be Honest with Your Bot

But now begs the question: How? How do we best use these systems and deploy them to the point where they don’t annoy web visitors and potential marketing prospects like Aaron Dun? For starters, be clear and honest that this is a virtual agent talking, said Ed Smith, chief product officer at Humley

You also need to have a clear and easy way to reach a live agent. “You need to make it very clear that it’s very possible to speak with live agents,” said Smith. Chatbot deployments should not be an exercise in trying to keep up customers at an arm's length, he added. “This isn't an exercise in trying to divert attention away from live agents,” Smith sadi. “That’s not going to serve anyone very well.”

Recognize your chatbot strengths, said Smith:

  • Fast responses
  • The ability to read what a visitor really wants
  • Ggoes away when necessary
  • Fulfills tasks when possible

Platforms Forcing Compliance

In the Facebook Messenger bot environment, companies who deploy chatbots have to be on their game — or risk the chance of Facebook shutting down the actual bot. Where Dun on his website was not able to “X out” his bot, users in Messenger can simply “block” a bot and never hear from it or that brand in Messenger again, according to Mike Balducci, vice president of strategy at Valassis Digital, which deploys chatbots. “In Messenger, as a consumer user, you always control that conversation,” Balducci said. “There are certain things that you're required to do as an app developer in terms of rules around when we can contact the user, what opt-ins we need for that user to be able to talk to them, what we can do when they don't reply. And then of course, the user can always block the conversation at any time.”

Facebook lays out its blocking-bots expectations and parameters in its policy overview. Companies must ensure their bots are “very user friendly and not overly pushy because if you get too many blocks on Messenger, Facebook is going to pull your bot.”

Related Article: The Next Generation of Chatbots: Connected, Intelligent, Capable

Be a Good Digital Shopkeeper

As for Dun, who called himself “frustrated in Boston” in his LinkedIn post based on his recent website chatbot experience, it comes back to the digital shopkeeper reference. If a salesperson in a store asked you five times if you need help, you’d “run away and literally leave.” Why can’t chatbots recognize that same logic? “If you design and you deploy your bot adopting the storekeeper metaphor, you have to adopt it all the way through,” Dun said. “If you don’t it becomes kind of counter-effective.”