We’ve all had that moment: You open a website or dial a customer service number and find yourself greeted by an obviously automated response. You curse the heavens and brace yourself for a “chat” with the soulless machine on the other end.
Of course, chatbots have made serious progress from their humble beginnings. The latest hotshot software, Cogito, can listen to human conversations between service representatives and customers and provide real-time feedback on how reps can adjust their conversation techniques. What was once a novelty is now a technology capable of providing legitimate quality-of-life improvements.
Despite these advances, chatbots still represent reactive technology, able to provide feedback when directed but still unable to proactively anticipate needs or work outside the traditional Q&A format.
Don’t get me wrong; they do have their advantages — after all, 20 to 25 percent of my company’s clients utilize chatbots in customer service. But before it can truly be considered useful, this AI has quite a bit to learn about human interaction.
June 5th: The AI Audit in NYC
Join us next week in NYC to engage with top executive leaders, delving into strategies for auditing AI models to ensure fairness, optimal performance, and ethical compliance across diverse organizations. Secure your attendance for this exclusive invite-only event.
The most valuable lessons for chatbots could come from the most unlikely of sources: professional pickup artists.
‘Picking up’ the lingo
Journalist Neil Strauss wrote The Game: Penetrating the Secret Society of Pickup Artists in 2005. While some of the tactics described are shaky, the concept of breaking conversations into systematic, guidable interactions could easily be applied to chatbot programming.
It starts with the opener, something to break the ice. Then follows the hook, which sets the conversation in motion based on the surrounding evidence. What do people want to talk about, what are the cues that indicate this, and how can the person (or chatbot) organize the conversation into a structured narrative with an objective in mind?
Next comes the exchange, which focuses on the possible outcomes of the conversation. For a pickup artist, this means getting a phone number (or not). For a chatbot, it means recognizing whether the customer is ready to purchase or might like more time to consider. Imagine a chatbot that not only places orders, but one that also recognizes hesitation and offers to follow up in a few days.
Finally, there’s a change in context: Where pickup artists try to go new places to further ingratiate themselves with their prospective dates, chatbots need to be plugged into multi-platform data. If the user moves from Twitter to the store page to Instagram, the chatbot should be able to recognize the person and use his or her history to build a more meaningful conversation.
Getting fake bots to have real conversations
If you currently use chatbots on your site (or want to take advantage of their capabilities), these tips will help you create more meaningful conversations:
1. Cater to the customer’s intent
If your customers typically contact chatbots for administrative purposes — changing a password, updating an address, etc. — the chatbot needs to be direct. If your customers tend to start a little broader, like asking for product recommendations, you can be looser with your chatbot approach. Understand why your customers use your chatbots so you can provide the experience they’re looking for.
Erica, the soon-to-be-released banking chatbot for Bank of America, is primed to be the perfect resource for customers in this regard. The company has been feeding massive amounts of customer data into the bot’s algorithms to prepare it for a variety of payment, transfer, and administration-related questions as well as preparing any more complicated issues customers may wish to address.
2. Create respectful rapport
Customers appreciate feeling recognized, but they might be wary of too much familiarity. Use past data points to let your customers know that you know who they are, but if you have more data than they might be comfortable discussing (such as browsing habits), don’t bring it up.
3. Let it flow (but keep it focused)
You want your customers to feel like they can do anything through the chatbot, but you still need to be able to guide the conversation through specific points to be productive. Think of it like the Grand Theft Auto series. The open-world format provides the illusion of endless possibilities, but there’s always a reminder of the next storyline to keep you moving forward. Give your customers the ability to perform different tasks, but keep them funneling toward the end goal.
4. Prepare conversation trees
What should that end goal be? That depends on the customer’s needs. Maybe it’s an update to a profile; maybe it’s a documentation request or a purchase. Whatever the need, the chatbot should be able to recognize that need and jump to a new path to help the customer along the way.
Take a look at H&M’s approach to this issue. Bot interactions begin with a quiz determining the customer’s style preferences — which also targets and focuses the bot’s potential recommendations. As a result, the software can cater to customer tastes without a human being pulling the strings.
5. Leave the tough stuff to humans — for now
Chatbots are excellent at relieving support staff of administrative tasks and helping customers fulfill simple requests, but the technology isn’t ready to replace the real customer service experience. Don’t fire all your service reps and replace them with robots tomorrow. However, if you have a simple task people perform often, chatbots are a great automation option. Domino’s now lets users order a pizza via Facebook by typing “PIZZA” to a chatbot. This accomplishes a simple task and frees up reps for more complex requests.
Chatbots might not be ready to debate politics with you, but they can be a great addition to your customer service force today. Use these tips to optimize your automated chats, free up your reps, and put some of the power back in your customers’ hands.
Henrik Werdelin is the founding partner of Prehype, a venture development firm.
