What are bots made of?. Not all chatbots are born the same… | by Joanna Trojak | Mar, 2021
Not all chatbots are born the same. Some of them are simple; some of them a little bit smarter. In this article, I will approach bots from the perspective of 5 Levels of Conversational AI.
Levels of conversational assistants have been introduced in the article The next generation of AI assistants in the enterprise by Alan Nichol, co-founder of Rasa. The report shows how to create a chatbot that is more accessible and user-friendly.
Levels are distinguished by how much responsibility falls on the user or developer. On the first two levels, the user has to work harder to get the chatbot’s value. The onboarding is essential to understand the way to converse with the AI assistant. With each level up, the user has less responsibility, but the developer has to work harder.
The chatbot sends simple notifications on your messaging app. It could be a notification about the cinema ticket reservation for Monday evening or a reminder about the necessary appointment.
The user has to know specific commands to get the information from the bot. Before Windows and Mac Os has become widely available, we were all using the command line to talk with the computer. This is very similar to interacting with the notification chatbot.
It is the most popular type of chatbot at the moment.
The chatbot has a predefined database of the most common questions regarding a specific topic. It is a slightly improved version of the FAQ site.
The problem is when the user wants to ask something that is not on the list or asks it in a way that is unknown to the chatbot. In a way, FAQ Assistants are just elevated versions of notification chatbots because you still need the specific commands to get the most out of such bot.
On the third level, the developer understands that the context matters and that day-to-day communication is not just a simple question and answer. It is essentially what the user said before.
The developer has more work to do because she has to implement the slots or other type of variables enabling the chatbot to remember important information like the user’s name, age or bank account number.
1. How Chatbots and Email Marketing Integration Can Help Your Business
2. Why Chatbots could be the next big thing for SMEs
3. My Journey into Conversation Design
4. Practical NLP for language learning
Here we are entering the level depicted in the movie Her. The more user interacts with the chatbot, the more chatbot understands him. It knows when is the best time to get in touch, what is your daily routine or what are your interests.
On this level, the chatbot understands emotions and what motivates the particular user. In the process, it becomes your friend, your constant companion and your assistant.
You may even fall in love with it.
On the last level, chatbots are fully autonomous and know every customer personally. Eventually, they can run large parts of the organization on their own.
This level will take years to achieve because the chatbot has to understand business and the people around it.
- Not all chatbots are the same. There is 5 step scale describing different types of chatbots and their involvement in our life.
- The first level — you only use simple commands, and the chatbot can send you simple notifications.
- The second level — chatbot has a superficial knowledge about FAQ, and it can only operate around those questions. It has a problem when the user does something unexpected.
- The third level — the bot starts to understand the context and act accordingly.
- The fourth level — the chatbot is personalized. It understands your personality and habits. It becomes your daily companion.
- The fifth level — the chatbot is autonomous, and it can run vast parts of the organization entirely on its own.
Credit: Source link