4 Subtle Ways to Make Your Chatbot More Human

The bots of yesteryear were not nearly as capable as the bots of today. Ask anyone on the street, and they’ll remember the “chatbots” from services like AIM and MSN Messenger. They could answer questions and return search results, similar to Alexa or Siri, but they weren’t impressive, at least not in a human way.

There was always the truth right there in front of you. You were talking to a computer programmed to answer and respond in a specific way. It was not a real, thinking, breathing person, advanced artificial intelligence or machine-learning system. If you went outside the bounds of the script the chatbot had been designed with, things would break.

But today’s chatbots are different, and Siri is a great example. Not only does she respond to conventional queries, but she can crack jokes, make suggestions and even help you perform certain actions, like make a calendar appointment or place a food order.

Siri seems surreal, almost human in a way. But that’s the trick, isn’t it? Before you can get on the same level as the greats like Alexa, Siri, Google Assistant and Cortana, you have to invest a lot of time and work. It’s not easy to make a chatbot or assistant seem human-like.

Searching for That Human Touch

Even with those aforementioned platforms, there seems to be something missing at times. Your conversations and interacts with those platforms are not ever truly human, are they? Subtle tones and nuances don’t always shine through. Sarcasm is delivered with a deadpan tone. Humor and wit are often lost on these “bots,” especially in the face of more complex jokes and conversations. They can make quips back, sure, but they’ll never be on the same level as a human friend or partner.

All that aside, there are ways to cheat and make your bot more human, giving it just enough character and personality to make it seem alive. This will help engage customers, who say they prefer to have a human helping them. Here are four ideas for improving your bot’s humanity.

1. Use a Real Voice for Your Chatbot

If your chatbot doesn’t use “audio” to convey messages, this rule still applies. The personality and voice of a person show just as well through text and the written word.

Consider for a moment what Alexa, Siri, Cortana and Google Assistant all have in common. A human, natural-sounding voice. They don’t sound like R2D2 or C3PO. Having a conversation with them is like talking to a friend or family member.

A lot of that is the way in which the assistant was designed to interact and function, yes, but it also has everything to do with their voice. The way they form sentences, read information and respond — it’s all human.

Give your bot a voice. Microsoft’s Tay, an AI-based Twitter bot, was remarkably human before things took a disastrous turn and she started spouting racist sentiments. Again, even though she was text-based, she still had a unique voice. It really makes all the difference.

2. Teach Your Chatbot to Display Emotion

No, your bot, AI program or computer are never going to truly feel emotion, at least not on the same level as a human. That said, there’s something visceral and endearing about a system that can react according to emotions and moods.

For example, if a bot chatting with someone detects anger, it could react with a sad or anxious emoji. For someone who seems sad or down, the bot could throw out an encouraging quote, uplifting message or just all-around friendly message. Imagine:

“Alexa, I’m not feeling too hot today.”

“I’m sorry, would you like me to order you some chicken noodle soup?”

It’s easy to look at an exchange like that and feel a human connection, even when there is none, and that will help your company. More than 80 percent of people say they will keep doing business with a company that provides a positive experience.

3. Branch Out With Your Chatbot’s Language

Yes, in some cases, it’s more of a hassle to include a contingency for every possibility. Sometimes, there’s only one direct way to answer a question or share information. Other times, however, there are multiple ways to respond.

That’s why you need to provide a “response bank” for your system that includes a variety of message options. If it just keeps repeating the same message over and over again in a loop, it becomes immediately apparent you’re talking to a preprogrammed computer or AI.

Branch out and provide varying responses for different situations. Your bot doesn’t always have to say “okay,” for instance, to confirm an interaction. You could have it say a variety of messages. Consider using “10–4,” “roger,” “confirmed,” “I’ll get right on that” or similar responses. Branch out, similar to how you would in game development.

4. Make Sure Your Chatbot Stays in Character

Ultimately, your bot has a purpose and a personality. Undoubtedly, when you sat down to give it a unique voice, you identified what these traits were. It’s important you take it a step further and never breach character.

Your bot — voice or text-based — should always communicate and respond in the same tone, same dialect and same manner. Personalities should never change. Moods should remain relatively even and predictable. Imagine Siri shifting her sophisticated manner by talking in slang. It would seem misplaced.

As soon as there’s a breach of character, a bot loses all its illusion of humanity. To keep your bot’s humanity in place, follow these tips and enjoy the improved response from your customers.

Image by Jack Moreh