5 tips for building the next great chatbot

Most chatbots today are a simple one-and-done. Out of curiosity and excitement, they draw a crowd. People play with them temporarily, only to disregard them minutes later.

Instead of being wowed or walking away in a state of awe, people are finding bots disappointing. The good news is that this represents a huge opportunity for those of us who can create a remarkable experience.

We are in the Wild West stage for chatbots. There is opportunity everywhere, and this can be overwhelming. There are a number of ways to narrow down the options and get ahead of the market. To have and gain clarity before everyone else does, and to have a meaningful competitive advantage, here are a couple of insights.

1. Follow the WeChat model

Bots have been on WeChat for a number of years. Their bot market is in a much more mature stage of development. There are many insights we can gain from researching what is working on WeChat and why. I have begun some research in this area and have a few takeaways:

1. Two types of bot accounts: WeChat offers Subscription Accounts and Service Accounts. Subscription accounts are the vast majority of accounts, and their main function is aggregating news and blogs. Service Accounts on the other hand are fully functional bots, with some superior functionality (like speech recognition). These accounts do things.
2. News: News bots dominated the Top 500 Bots in China. In fact they make up about 34 percent of all bots in top 500. WeChat serves as the main way most Chinese people get their news.
3. The top five most popular categories: Current Affairs, Culture, Entertainment & Sports, Humor, and Emotions — these five categories made about half of the top 500 bots on WeChat in July 2016. Entertainment & Sports and Humor show to be good use cases for bots.

It is important to remember that there are strong cultural components at play. China has a poor app culture, mostly due to low-quality smartphones and cost of data. As a result, bots have been the main way to go. Moreover, information in China must be government approved. There is no Google and no free press, and as a result most Chinese people are getting their news via WeChat.

(The above information about WeChat was translated and summarized from this link by Tianyu Guo.)

2. Focus on chatbot-only use cases

Every new technology inherently brings new ways of doing things that were not possible before. As a result, there are new use cases, new ways to solve old problems that are made possible as a result of this new technology.

So here we have to consider the question: What problems can you solve with chatbots that could not be solved otherwise?

Part of the reason Uber has been so successful is because of this strategy. They leveraged new technology (smartphones with GPS and payment capabilities) to do something that could not have been done before, namely ride-hailing.

Currently, bot makers have the same challenge.

Examples of chatbot-only use cases:

Legal Help: DoNotPay has overturned 160,000 tickets in London and NYC, saving customers over $3 million. The bot, which takes 25 percent of the savings, has made nearly a million dollars.

Customer Service: Right now, customer service is a poor, poor experience. No one wants to call an 800 number, and most companies dread having to improve their customer service. Bots will change all of that.

Personal Advice: Get real advice from real humans. For example, the Sensay bot connects you with “people” who can help right now via chat.

3. Make it unique

Chatbots are a natural expression and an externalization of what we do almost every moment of every day. That is: We talk.

An average person has between 50,000 and 70,000 thoughts per day. We are constantly thinking; in other words, we are in deep conversations with ourselves.

Words and conversations are the human operating system. For the first time, there will be a direct overlap between the Human OS and Machine OS, between how we think and our technology. Soon we will be able to talk with A.I. as though we are talking to our assistant. Or when we have a bad day, we can brood about it or talk to an A.I. bot instead.

A.I. is not there yet. It is not even very close. However, between here and there, there is opportunity in niche spaces where conversations are needed in order to solve problems. Bots that can solve problems in this way will have the ideal form and function and be able to help us get things done in the way we know best — via conversations.

4. Tap into the real need

Most of our thoughts are completely centered on our fears, our desires, our problems, and our needs. One of the major problems in our modern world is that many of us are feeling lonely, disconnected, and even forgotten. We are constantly looking to connect with the outside world.

A recent survey of over 12,000 users of Assistant.ai revealed that nearly 40 percent could imagine themselves falling in love with A.I. and another 25 percent agree that “it could happen” to them.

Humans have a strong tendency to anthropomorphize and assume relationships. This is why we see faces in the clouds, why Brian (the dog in Family Guy) walks and talks like a human, and why we transpose human characteristics on the family goldfish and begin to have a relationship with it. We are wired to connect and are constantly looking to relate.

Many humans are yearning for connections so much so that even poorly made chatbots offer us the opportunity to anthropomorphize and assume a relationship.

This tendency is so strong that people in China and Japan have spent so much money on virtual girlfriends/boyfriends and some have even married their virtual sweetheart.

5. Hit a nerve

In order to get people to care about anything, to really care, you have to hit a nerve. You have to touch them emotionally, you have to make things meaningful for them.

The best way to touch someone emotionally and to inspire them to actions is via personality. Personality allows us to more easily connect and create a relationship, and bots are an ideal vehicle to a personalized marketing relationship.

They allow us the unique possibility of having a deeply personalized, intimate, unique experience with A.I. that meets our needs, both physical (by doing things for us) and emotional.

Putting it all together

All of this is available to us today! Our goal when creating the next great chatbot should be to leverage the uniqueness of the platform and solve problems that could not be solved otherwise. In the process of solving these problems for people, our bots should be fun, engaging, positive, have a lot of personality, and also give people love.