Chatbots are hot right now. They’re rapidly changing from just a buzzword within the creative and marketing industries into a real emerging marketplace and new interface with the potential to radically alter the mobile landscape-—and the mobile moments that occur between brand and consumer.

If this is all new to you, don’t worry. A chatbot is defined in the Oxford English Dictionary as “A computer program designed to simulate conversation with human users, especially over the Internet.” The dictionary tellingly adds, “chatbots often treat conversations like they’re a game of tennis: talk, reply, talk, reply.” Users interact with a chatbot via mobile messengers (such as Skype or Facebook’s Messenger), text messaging, or the web, with artificial intelligence (AI) typically powering the experience to enable context to be added to the conversation.

Why Chatbots Are Big News

Messaging apps on mobile are massively popular, having overtaken social networking apps for monthly active users (MAUs) last year, according to Statista. Currently, WhatsApp has more than 1.2 billion MAUs and so does Facebook Messenger. Keeping an eye on this space is a wise move for brands, since it is where your audience is likely spending its time. Even Amazon, Facebook, Google, and Microsoft are betting that chatbots will be big.

Bots Built on AI

The emergence of AI has been a work in progress to say the least, but it’s the trend this year. And the fact that AI is still developing hasn’t stopped large brands, designers, and indie developers from getting excited about them and the platforms that are being built for them. Conversation is allowing us to have more natural interactions with brands, and the AI behind them could end up being so sophisticated that chatbots are able to pass the Turing test.

So why are so many bots still failing their creators and the brands that pay for them? I’ve taken a look at seven of the reasons chatbots fail, so you can avoid the same pitfalls.

  1. AI is not quite there yet—AI and natural language processing (NLP) are not at a point at which they can be used to create products at a relatively quick and inexpensive rate. Whereas products such as Amazon Echo and Google Home are built on NLP, a vast number of chatbots are being built on decision-tree logic, in which the response a bot gives depends on keywords identified in the user’s input. The user is taken through a route of conversation based on what he or she has asked. Unlike Alexa, these chatbots are limited by the number of variables the development teams have allowed for.
  2. Bots cannot handle us—Bots cannot handle human language and context in the same way that people can. Language as a means of communication has been under development for millennia. The subtle nuances of tone, slang, facial expression, and even sarcasm take us many years to read properly. We’re constantly reading between the lines. Expecting technology in its infancy to be as sophisticated as thousands of years of evolution is unrealistic.
  3. Use cases are still not strong—Many brands have attempted to build chatbots in order to beat their competitors to the punch. This means they are not built with particularly strong user needs in mind. They were created because they could be created, not because they should have been created. Often, messenger bots are not saving people a lot of time—they’re just offering them a different way of getting to the same goal. When technology and products aren’t built based on user needs, they tend to frustrate and offer poor user experiences.
  4. Mass adoption hasn’t happened—Early data shows that the younger you are, the more likely you are to interact with chatbots. This is because young people are most likely to adapt their behavior. Older folks haven’t warmed to chatbots because they don’t like change and have set digital and mobile behaviors.
  5. Chatbots haven’t found a home—Some chatbots function as their own apps. Some leverage pre-existing experiences. Others are built into websites and platforms. Chatbots should exist at the point in an experience in which there is some benefit. However, since that point can differ so widely, there isn’t a standardization of patterns—which means they can be hard to find.
  6. Quantity over quality—Since chatbots are cost-effective and fairly easy to build, there are too many of them. It’s hard to separate what is useful from what is not. Our desire to build them, twinned with many aimless executions, is tarnishing their reputation. It can be argued that there are more bots than people using them. If you want to stand out, use your chatbot to solve an actual problem.
  7. Bots and brand ecosystems—Bots also tend to lack escalation processes. When you call your bank and can’t get what you need from the automated system, you’ll be connected to a human. This is not necessarily true when you use chatbots. The result of this is frustration on the part of your users. Don’t forget the human touch—people want quality of customer experience.

The chatbot market is still cluttered, but there are some glimmers of inspiration coming through. Many are changing the way we interact with certain brands, such as figuring out where to go on holiday or replacing poor on-boarding experiences of website forms. Brands are getting it right. ?The integration of chatbots into the mainstream of how we live our lives is dependent on the AI technology they’re built on. As machine learning gets stronger, so will they. And as machine learning’s learning curve becomes smaller, they will be harnessed and used for more powerful reasons. The chatbot journey is just beginning.