Perfect Bots from Hell
For one week we created and launched one chatbot per day on Facebook Messenger. All of them are pretty stupid, but we still learned something from it.
At K2 agency we are a great believers that conversational interfaces are the future and will be very important in the coming years. Just check our work for a very serious industry— banking.
But we also like to have fun at work. We believe play is the highest form of research. So one day at lunch, at a Chinese restaurant near our office (Chicken Hanoi is really good!), we came up with one crazy concept.
“Let’s make a bot, or rather a whole family of them (we created a dozen of concepts), that will be completely absurd, useless, frustrating, and stupid. But they will also be perfect, because they will always deliver at 100%!”
The biggest fear to employ chatbots is that they will not be able to understand all the questions asked by the users. You know, this scenario:
But it would never be the case with our bots. They will be designed to always give the perfect answer (by their own rules)! 😝
By the end of our lunch we had 3 big principles for our bots:
they should always answer correctly (in their own way) for every question
they should be very easy and quick to make (though we had to scratch our heads a few times even with something so simple)
and they should be stupid (which goes well with point #2!), but somehow clever at the same time, and funny (for us at least)
So we built one bot per day, according to these principles, launched them on Messenger, and had fun observing people interacting with them.
Before I introduce you to our happy family of stupid bots (and then you will understand better), let me explain why it was worth for us to do something like this?