How to Make Your Chatbot More Human?

Apr 17, 2019 · 8 min read

Chatbot development cornerstone was laid in 1966 by Joseph Weizenbaum. The principles used in ELIZA, the first chatbot ever, was a foundation for the development structures of chatbots we know today.

Fast forward to 50 years later where we are witnessing the skyrocketing number of chatbots around the world. Some of them live up to the modern standard of NLP or smartbots, while others are still following the simple principles ELIZA had, such as keywords, specific phrases, and preprogrammed answers.

In a crowded chatbot race, it is essential to develop a chatbot with a smile to stand out! And by a smile, we are thinking about the BOT personality which will give it a sense of realness.

Advanced Chatbots and Human-like Characteristics

Today’s chatbots are advanced. Advanced and contextual chatbots utilize Machine Learning (ML) and Artificial Intelligence (AI). It makes them remember conversations with specific users to learn and grow over time. Unlike keyword recognition-based chatbots, they can self-improve. They can learn from what users are asking for and how they are asking it.

For example, a contextual chatbot that allows users to order pizza will store the data from conversations. They will learn what the user likes to order. The result is remembering the most common order, delivery address, and payment information. Instead of having to respond to several questions the user just has to answer with ‘Yes’ and pizza is on its way!

Related: Emotional Intelligence — The Future of Chatbots

Example of chatbots like that is Siri. She can crack jokes, make suggestions and help you perform certain actions. She can make a calendar appointment or place a food order. Another one is Amazon’s Alexa. She can inform about the weather, stream news and music on demand and serve as a robotic assistant. She responds to voice commands to control home lighting and much more. There is a lot of similar ones such as Microsoft’s Cortana, Facebook’s M and the unreleased Viv.

Empathic Virtual Agents

However, only Al assistance is not enough. According to the case of Ellie, people tend to have an emotional connection with the interlocutor. Ellie is an AI psychologist that has primarily been used to treat military personnel suffering from PTSD. She uses verbal and nonverbal cues and engages in a conversation. Interesting about her is that patients prefer talking to Ellie over a person.

People didn’t feel judged and were provided with more information. This is important to understand what kind of assistance people feel comfortable with. All that aside, there are ways to make your bot more human. By giving it enough character and personality you can make it seem more alive. This will help engage customers who prefer to have human assistance helping them.

A Real Voice for Your Chatbot

Voice search has an increasing role in how people look for content. According to BrightLocal Study, voice search popularity is growing fast.

58% of consumers have used voice search for local business information within the last year
46% of voice search users look for a local business daily
27% visit the website of a local business after conducting a voice search
76% of smart home speaker users conduct local searches at least once a week
— with 53% performing daily searches

Voice search revolutionizes the way people look for the product, service, and information.

Because it is simple and much faster, voice search will take an even bigger role in searching for content. That’s why it’s a good idea to start adjusting your marketing tools and integrate a smart voice chatbot. Except for the ability to “listen” and respond in voice, the type of voice it projects also matters.

Advanced bots such as Alexa, Siri, Cortana, and Google Assistant have in common natural-sounding voice. They don’t sound like robots, R2D2 or C3PO. They respond like a friend or family member would. A lot of that is how assistant interact and function. Also, it has everything to do with their voice. From the way they make sentences, read information and respond — it is all natural and human.

For example, Microsoft’s Tay, an AI-based Twitter bot had human characteristics before it took a disastrous turn. She was a text-based bot but with a unique voice. Although this is just a small detail, it really makes a big difference. Likewise, Duolingo has developed voice chatbot that helps people learn a new language. They have noticed the hardest thing with learning language is practicing loud, so they provide their customers work on spoken language.

Picture 1. Duolingo Chatbot

Make Your Chatbot Emotional

People prefer talking with someone who they can identify. It is easy to look at an exchange like that and feel a human connection. Even though they are aware there is no human, it feels better when the other side shows understanding and emotion.

Consequently, more than 80 percent of people say they will keep doing business with a company that provides a positive experience. You can apply the same with bots who arouse positive feelings.

Here emotional API can help. Emotion Application Programming Interface (API) enables you to recognize emotions in images. Emotion detector takes a facial expression and returns a set of emotions for each face. It can detect a range of emotions like anger, contempt, disgust, fear, happiness, neutral, sadness and surprise. These emotions are understood to be cross-culturally and universally communicated with particular facial expressions.

To get an image of facial expressions you can use TrackingJS. It analyzes a video and it is based on the configuration that we have given it. It also tracks for things like objects and colors. When finally chatbot notice and understand emotion, you can make him respond more friendly.

For example, if a user types in: “Worst day ever. I arrived late because of the traffic.” A chatbot can say “Sometimes life just sucks!” in disgust mode.
Also, it can answer in other moods such as liking “I am always here to support you.” Further, “Keep smiling! Things will get better.” is an example of a happy mood.

Stand out with Your Chatbot’s Language

The same as voice and emotions, your chatbots language can also make a big difference. Although for some questions there is only one direct way to answer them, sometimes you can differ. Sometimes there are multiple ways to respond. That’s why you need to provide a stock of responses. If your bot keeps repeating the same message, it becomes apparent that he is not a human.

There are chatbots in specific industries that should have adjusted language. For example, Amiya is Dressipi’s fashion assistant that smoothens the shopping journey for the customers. Fashion chatbots provide customers with a suggested outfit and also gives size recommendations, in real time. They use language enriched with colors and fashion terms.

Traveling chatbots can book and schedule flights, and help integrate additional services via social media platforms. One of them is American Express bot Mezi. Typical language for Mezi is “Any particular part of the city you would like to stay?” and friendy travel suggestions.

Healthcare bots can determine symptoms and suggest medicine.

They can use sympathetical language like “I’m sorry to hear that, Anna. How have your symptoms changed, if at all?”

Picture 2. Style Adviser by Dressipi

Define Your Chatbot’s Character

Ultimately, after giving your bot voice, emotion and language, it needs character. A character is one step further to a more realistic personality. Your bot either it is a voice or text-based, should always communicate and respond in the same tone. It includes the same dialect and same manner.

If you imagine Siri shifting her sophisticated manner by talking in slang, you could understand how inappropriate it is. To keep your bot relevant to the target customer, their character should be adjusted. Diversity of perspectives, backgrounds, and skills is key in personality building. When making bot’s personality, you also need someone to embody the target user. Make bot similar to your target customer.

There are 12 archetypes commonly used to define a brand persona. There are many tests and visual representations that can help you and your clients determine what brand persona your chatbot should express. Here’s one good example:

Picture 3. Chatbot Personas

The archetypes are a starting point. You’ll still need to fill in the details of your persona. When developing a chatbot based on an individual, knowing that individual’s archetype and the way they express that archetype in their written and spoken communications, can help you develop new content.

Also, after you choose a character, you need to choose the way it will deal with objections. Where there are sales, there are objections. Your chatbot will deal with them, and there is a way to help it. Here is important to have data but before you get it you can use assumptions. After you get data, you can start tweaking your sequence in order to better overcome users’ objections.

Demographic Characteristics

The last thing is to give your chatbot some demographic characteristic. First of all, a name would help easy to remember him or her.

After that, determine bots age the same as your buying persona is. That will help your customer feel like they are talking to someone who understands and cares about their needs and pain points.

Second, have in mind that localized content can be extremely effective in marketing. Knowing that you can localize the content of your sequence. Further, you can give your bot place of living. Also, it can have its favorite quote and goals, the same as values.


Since there are many ways to personalize chatbot, it is understandable that chatbots will continue to look more like a human. Natural sounding voice will encourage a customer to “talk” with your bot.

Your bots personality and character make it easier to relate to the customer’s problem and finding a solution. And finally, the bot’s voice makes everything more realistic and faster.

Many sources show that messaging apps are on the rise. It is natural that if they are growing, the need for conversational experiences within is constantly increasing. According to a study by Statista, by 2021 there will be 2.5 billion messaging users worldwide.

Thus, as it is obvious, the more the chatbots, the more difficult for you to point out with youts bot. Chatbots will more and more resemble human and it will we hard to differentiate them from a real human. This is one more reason why the personality of your little friend is very important. It helps it to stand out from the crowd and makes people want to explore it, rather than avoid it.


Written by


Omnichannel communication platform that helps businesses get in touch with their future customers.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade