When Philip K Dick wrote "Do Androids Dream of Electric Sheep?" in 1968, he imagined a future where artificial beings are almost indistinguishable from humans. They look, sound and act alike — and mimic human emotion and intelligence.
What we saw emerge in 2016 with the launch of Facebook’s Bots on Messenger initiative was very far from that.
Chatbots are implementations of conversational user interfaces and typically live in messaging environments. They attempt to converse in natural language, but often feel quite rigid, and sometimes outright dumb, repeating “I’m sorry, I didn’t get that” more often than they should.
So are you free to ignore chatbots? Far from it.
Can We Talk?
Despite the failed expectations, which were far too high to begin with — and influential sci-fi stories are partly to blame for those high expectations — chatbots can prove invaluable for the Chief Marketing Officer and Chief Customer Officer alike, if done right.
Anyone engaging with customers wants to do that in the most, well, engaging way. Being able to meet customers where they are (with billions of users already using the likes of Messenger, WhatsApp or WeChat) is promising. Though still clunky at times, chatbots converse with customers in a more natural way than previous digital solutions — and they're getting better.
We’ve been trained for decades to click on menus and icons as the best way to engage with a computer and, as a result, with companies. Oftentimes it’s a guessing game trying to make sense of the options we see or hear, and sometimes we have to backtrack to correct a wrong menu selection.
Why? As humans, we have spent our lifetimes requesting information, sharing ideas and more through conversation.
Recent advancements in speech recognition over longer distances have brought about the fascinating home automation and assistant devices from Amazon (Alexa on Amazon Echo) and Google (Google Home).
We are on the verge of a new paradigm shift in man-machine interaction. So if the technology has indeed advanced and UX designers and developers are both flocking to work on this new trend, how can a customer-facing organization take advantage?
Anticipate Customer Needs
It comes down to determining which questions the customer asks at the various stages of their customer journey:
Demand and Research: “Do you offer life insurance?” “I’m looking for a cordless water boiler” “Do you sell the new iPhone?”
Product Inquiry: “What are your rates to ensure my new Audi A4?” “Do you have the K570 in stock at your Cambridge store?” “What is your return policy?”
Product Sale: “I’d like to buy the K570 in red” “Can I purchase the iPhone 7S+ in jet black?”
Service and Support: “Where is my order?” “You charged me twice, can you fix that?” “I’d like to re-order the hand lotion, same delivery address as last time”
Surveys: Now it’s you asking them: “How would you rate your service experience?” “What did you like best?” “On a scale of 1 to 5, how convenient was the check-out process for you?”
Not all stages lend themselves equally well to being served with a chatbot, and don’t think a bot can replace any of your existing channels.
Bots are about additional convenience. If the customer has a question and it is straightforward, a messaging interface allows them to ask it in the most efficient and direct way possible.
The more your customers get used to the different service channels you offer, the more they learn which is best for their current need and situation. If they’re driving, the phone channel is the only option. If they’re mobile, apps or bots can be most efficient. A full discovery process for a new purchase might best be done on a full-blown website at home or at work — but it might still start with a question on a messaging channel, then carry over to a richer environment.
An Architecture to Support the Flow of Conversation
As you think through the questions your customers have in the different stages, you are already preparing the so-called Conversational Architecture (CA) work needed to build a chatbot that can help with these questions.
The CA lays out what content the bot can expose, which answers it can give. It leads to the design of the dialog flow and the concrete messages the bot will use when conversing with your customers. After the implementation and rollout phase, you will want to carefully monitor the use of your bot, to adjust it for real-world needs and make it better over time — just like you would with new hires in your contact center or sales departments!
Start Small, Dream Big
The key to a compelling experience on the messaging channel lies in three things: 1) spending enough time on the CA/flow/message design of your chatbot, 2) setting the right expectations with your customers as they learn what this new channel can do and 3) human backup at all stages. A deep integration with your contact center is something you cannot do without, just like you would never leave an unseasoned agent or sales rep on the floor of your brick and mortar store alone without subject matter experts or supervisors ready to jump in.
The good news: with the human backup in place, chatbots allow you to start small, and that’s what you should be doing.
If you follow this advice, invest enough into the design of the experience leveraging the expertise of the right vendors and don’t overestimate what the technology can do, you just might end up delighting your customers through innovation.
And who knows, maybe one day YOUR chatbot will dream of electric sheep.