A bot in a chat app is not automatically a chatbot

If you look into chatbots a little, you soon find that chatbot is a hard term to pin down. It is often confused with the term bot, while a bot and a chatbot are anything but the same. For many people the definition of the word chatbot is therefore not clear.

If in an app such as Facebook Messenger you are presented with a menu of choices, that is called a chatbot. And the simple little bots with which ten years ago you could look up your bank balance in MSN Messenger were already regularly called ‘chatbots’ back then.

However: a chatbot is only a chatbot when you can actually hold a free conversation with it. With Watermelon's chatbot a user starts a conversation and can in principle ask any kind of question they could also put to a customer service. They do this in normal human language and the chatbot answers in normal human language. An important part of that is that the chatbot is not directive. Of course it indicates when it does not understand your question, but it does not steer you as a user in a particular direction. It does not go through a standard script, but holds a conversation thanks to artificial intelligence. That is something we have never seen with bots, because it was simply not yet technically possible.

‘Normal’ bots

That makes a chatbot, such as Watermelon's, not only much more advanced, but above all more versatile. Which does not mean, incidentally, that there is anything wrong with a ‘normal’ bot; those can be very useful indeed.

The recruitment agency Randstad, for example, has a bot that helps you look for a job. Instead of entering all sorts of search criteria on a website, you simply start a conversation with Randstad on Facebook Messenger. The bot, called Vera, then asks you what kind of work you are looking for, how many hours you want to work, which region you want to work in and how much you want to earn.

On the basis of your answers the bot searches Randstad's vacancy database for suitable vacancies and shows them to you. A very handy bot, because it has a very clear and useful purpose. There is nothing wrong with that at all. Because your search here feels like a conversation in normal human language and takes place in a chat app you already use anyway, it becomes even easier for people to look for a job.

This bot is not a chatbot, though. You cannot really talk to it. Vera starts straight away with a script and asks every user the same kind of questions, in the same order. On top of that the answer options are usually limited to a number of choices. As a user you never really type a whole sentence, let alone several sentences; you only give answers to questions. This could also have been worked into a wizard or form, which incidentally does not detract from the fact that it works very pleasantly for users.

Randstad's Vera is of course only one example. Another example is the business chat app Slack, which can be expanded with all sorts of bots. With one simple command you can have a bot search for a funny GIF that you share with your colleagues – at many companies the most used bot in Slack. Or you can instruct a bot to send a colleague a reminder at a particular time.

Customer service

Services in which as a user you give a simple command or as a maker you want a user to go through a number of steps are easy to capture in a bot. If you want to automate your customer service, though, that is not going to work that way. Because users often have countless questions, a full chatbot is the best solution. Via artificial intelligence the bot understands which question the user has and can answer it. And if it cannot answer a question after all, it can learn that for the moments when it is asked again in future.

This is exactly the reason that Watermelon developed its chatbot technology. Because a conversation with customer service cannot be captured in a script for a bot, but can be automated by means of a full chatbot.

A chatbot is therefore, technically speaking, a bit more complicated than a simple bot. Where with a bot you can still enter all users' commands or questions manually, with a chatbot that is not possible. People can ask so many different questions and do so in so many different ways. Watermelon takes the complicated work off its customers' hands, though. The chatbot's artificial intelligence, which actually understands what a user is asking, was developed by Watermelon itself and can be applied to any customer service. The best thing of all is that a user does not even have to notice that they are talking to a chatbot, so human is the chatbot's response. And that is something a ‘normal’ bot will never manage!

Would you like to know more about how to deploy chatbots successfully? Download the free whitepaper by clicking the button below!