The ins and outs about the modern digital assistant
| 4 min
It’s all in the name: a (ro)bot that your customers can chat with. In the past few years, the technology around chatbots has grown enormously and it is becoming an increasingly important part of your service. In this blog we'll take a closer look at the evolution of the chatbot, the different types of chatbots and we’ll give some successful examples.
The evolution of the chatbot
Chatbots are certainly not new. In 1964, the first bot named ELIZA was developed in the Artificial Intelligence Laboratory by American computer scientist Joseph Weizenbaum. Joseph wanted to demonstrate that communication between humans and machines was indeed possible. Currently, ELIZA can still be used. Despite the fact that the chatbot was not very intelligent, users still found that ELIZA sometimes sounded like a human.
The American company IBM came up with Watson in 2006: a supercomputer that could interpret a question and answer it within a few seconds with the help of encyclopedias, books, magazines, articles and websites. Two years later, Apple launched Siri; even though it is not officially labeled as a chatbot, it was a breakthrough. In 2015, its competitors Cortana and Alexa followed.
The start of AI chatbots
Starting in 2017, the term "AI chatbot" came along more often. While this sounds trendy, it initially caused a lot of frustrations by its end-users. Whereas the rule-based chatbot gives the customer options to choose from, the AI chatbot works based on open-ended responses given by the customer and responds with an answer that is most appropriate. The technology was still in its infancy in 2017, which often led to wrong answers. But developments have continued and today AI chatbots are performing really well.
How and when do you use a chatbot?
Customers are more demanding than ever. By implementing a chatbot, you can ensure that your customers receive their answers 24/7. Your new digital assistant gives the rest of the team more time to breathe. For example, a chatbot can be deployed to take over a part of the customer service so that most frequently asked questions are answered immediately. But a chatbot can also be used for lead generation or even within HR and recruitment. In the chatbot world we distinguish two types: an AI chatbot and a rule-based chatbot. We explain the difference between these two below.
What is the difference between an AI chatbot and a rule-based chatbot?
An AI chatbot
One of the main differences between an AI chatbot and a rule-based chatbot is that an AI chatbot actually learns to understand the questions customers ask. In this way, customers get a natural and personalized experience. This is achieved with the help of Natural Language Processing.
Natural Language Processing is a form of AI where the computer gets better and better at recognizing words, interpreting sentences and understanding the questions. In the case of a chatbot, this means that instead of choosing from predefined options, customers can simply ask any question and the chatbot will respond accurately. The chatbot learns to recognize different questions, spelling mistakes and synonyms. As a result, your AI chatbot understands the question and gives customers the right answer.
A rule-based chatbot
This chatbot offers the user a choice of options. This is done using a menu or buttons. Depending on what the user clicks on, the chatbot then provides a different set of options to choose from in order to give the right answer.
This structure is quite simple and is therefore commonly used. These chatbots can answer predefined questions and navigate users through a website or webshop, making their buying journey easier. These chatbots only quickly become ineffective when it comes to solving complex requests with many variables. So once the user's questions don't fit into the pre-designed answers, this type of chatbot can't help, which ends up being quite disappointing and frustrating for the user.
The true power of a successful digital assistant lies in the combination between an AI chatbot and a rule-based chatbot. We call this a hybrid chatbot. By uploading the most frequently asked questions into your chatbot, the vast majority of your customers can be helped within seconds. Customers with more complex or unique questions can simply type what they are looking for and the Conversational AI will work for them to provide the right answers.
But what if your chatbot doesn't know the answer to a question (yet)? That's when your chatbot becomes a real part of your service team. These questions are automatically forwarded to an available colleague, while the unknown question is saved so that it can be added to the chatbot later. The colleague receives a complete overview of the conversation and the customer so that he or she can immediately pick up where the chatbot left off. In this way, your customers will experience excellent and fast service and your team will have more time to focus on complex questions.
Chatbot examples from practice
Reduce workload with chatbot Kittie
Before Kitcentrum had launched chatbot Kittie, they answered every customer question manually. They grew tremendously quickly and so did the incoming questions. To answer all these questions everyday was almost a fulltime job. And so Kitcentrum decided to look for a better way to automate their customer service. Nowadays 85% of their customer questions are automated. Quite an impressive number!