How to Sound Human in the Age of Chatbots
The digitally conscious customer of today needs immediate responses. A key factor in staying ahead of the competition is the ability to provide proactive customer support to your customers. Live chat support came to the rescue and enabled agents to respond to customers as swiftly as they could. But this was not good enough. It is almost impossible for live agents to be available all the time for the omnipresent customer. This has opened the case for customer support enabled by artificial intelligence (AI): chatbots.
Chatbots are great! They help you proactively engage with customers, and they deflect basic how-to questions, allowing agents to focus on more complex issues. They help save every support agent’s time and improves their productivity. They can also be customized to be an extension of your company’s brand.
But chatbots are only as intelligent as they are designed to be. Their language processing module can make severe mistakes if they are not mapped to understand sarcasm, read between the lines, or get the context of the dialogue.
So how do you build a chatbot that can provide satisfactory interactions, while also being resourceful? The trick lies in building a chatbot that can be polite, intelligent, and viable. In this article, we have curated a checklist to help your chatbot attain such qualities, and more.
#1 Name your Chatbot
Naming your chatbot gives it a personality and defines its role. Once you finalize your bot’s capabilities, decide on a name that is catchy, relevant, and reflects its purpose. Avoid names that are too technical or boring.
For example – bot names such as SLA3000 and Shipping Tracker Bot sets the whole tone of the conversation to be very computerized and robotic. Giving your bot a human name is the first step towards humanizing it.
Initiating conversations by introducing bots with their names is also a great conversation starter, and keeps the conversation true to its motive.
When Idealist created Ida, a chatbot that assisted people with donating or volunteering for charities, they wanted a name that was short, catchy and easy to remember. Ida not only checked all the above boxes but also sounded similar to their companies name. Ida also means “to help” in Arabic, giving more meaning and context for the name. Moreover, “Ida for good” made a great phrase/hashtag to promote the chatbot.
#2 Greet Right
Chatbot greetings are a perfect place to set clear expectations on how your chatbot will assist. Greetings can also be scripted in a way that leads the customer towards taking at least one potential action. A friendly proactive greeting adds personality to your chatbot and creates a reliable impression, making the customer build an emotional connect with the chatbot.
Take a look at how Jeeves makes the first impression with its greeting. The bot not only conveyed its services but also offered options to choose from. This cuts down resolution time, sets defined assistance route, and ensures quality assistance.
#3 Be Transparent
A customer support chatbot’s primary goal is to assist people and direct them to what they want in a quick and accurate manner. Though they are motivated to mimic human behavior, bots cannot recreate human assistance, and can become clueless if conversations get deviated from the topic they are simulated to assist. Thus, introducing your bot as a bot in the first place sets the expectations more realistically, and your customers will not feel let down to realize they were being assisted by a bot and not a slightly mechanical sounding human being.
Try to stick to something like this example of ASAPmaps bot.
#4 Emoji-fy your Chatbots
Back in 2015, every app that allowed messaging started to integrate stickers, emojis and gif options in their keyboards, and there has been no looking back since. These animated features have made conversing more reaction oriented and has given a face to chats. And it’s only right that your bots make the most of this trend.
A UX advanced chatbot is expected to read these animations and reply with the same. For instance, chatbot Mica reads the emojis sent and replies with the list of hotels that serves the dish. Additionally, the chatbot can read audio files and reply in turn with audio files.
#5 Sprinkle Some Humor
Humour is not a trait chatbots are obliged to adopt. However, adding humor to your assistance shows that you are going the extra mile. Explaining mundane things with a pinch of humor not only grabs attention, it also retains attention. Humour also gives personality to your bot and gives it a reliable image.
It’s also crucial to understand how humor works. Humour has to be localized – meaning the bot should be designed to understand the context of the conversation, while also identifying when and where to crack an appropriate joke.
Meekan, in the above conversation, informs the customer that they might have to wait a couple of minutes. It was a smart move at the right time, because it gives a window for ensuring that the customer does not get restless and leave during the wait. ]
Moreover, mistakes are human. When your chatbot stumbles along the way, it can easily recover with a witty apology. People have not only dismissed mistakes in such instances, but have also appreciated brands for their comical approach.
#6 Use the Chat Lingo
English alters its way into new mediums that cannot adapt old-form languages. When millennials started communicating through chats, words got trimmed, and phrases got abbreviated. No matter how many thesauri your chatbot is routed to, without the context of the chat lingo, they won’t be able to do much. Make sure your chatbot doesn’t make the same mistake as the weather zone bot below.
#7 Understand the Human Way of Thinking
Context matters. People tend to break sentences in chats and correct them, without implying that they made a mistake in their previous statement. Customers can also take a satirical, sarcastic or cynical route to reply at times. A chatbot should be trained to navigate itself through the context of the conversation, so it can guide customers properly.
In the below chat, when the customer tried to modify what they first stated, the chatbot mistook it for 2 different orders and added both the items. Machines have a different way of understanding things than human. Try to make sure your chatbot does not make such mistakes.
#8 Stick to Simple Questions
Chatbots are built to be inherently social and assist human needs. Keeping your bot true to its conception and not complicate how they work is one way to ensure your chatbot meets customer expectations.
This principle works the other way too. Copy write your chatbot conversations in English, and breakdown instructions for better customer understanding. The following mishap could have been avoided if the customer was provided with the form, rather than asking them to go to the form through a chatbot.
There are a million chatbots popping up each day to serve us one way or the other, but only a few succeed. Fast learning artificial intelligence (AI) chatbots like Freddy are a huge hit in the market due to the fact that they benefit both customers and businesses. They are easy to set up for brands, engage customers through targeted campaigns, and provide 24/7 assistance at a lesser cost while being less invasive and more reliable for the customers. These successful chatbots also have another thing in common. They tick all the boxes in the checklist above, and have a human tone to them.