Prefer bots over fake humans

KLM will create bots to human interaction through Facebook Messenger. Microsoft will invest in helping developers build more bots. Slack is the home of many little bots that should help us things get done. Siri, Cortona will be our life assistant right at the fingertip.


There is plenty of evidence that more interaction between services and customers will happen through bots. As service designers, programmers and hackers it’s our job to make these interactions meaningful. But not only that. We should create great experiences out of it and be aware of the limit of the bots. After all the interactions I had with bots in the last months I got to the following conclusion:

I prefer a clear bot than a fake human.

Some companies try to create bots which act like little fake humans. They have names. They make little jokes. Sometimes people interacting with those do not even know that they are interacting with a bot. That is, at least for now dumb. Why? Because there will always arrive a moment where the bot, as smart as it is, will fail. And a bot which fails at pretending to be human is way more bad than a human failing at delivering good service. When the bots that we thought were a human fail and showed us its real nature. We feel fucked. We feel manipulated. «OK, this company doesn’t care about humans. They just want to automate the interaction with me! Screw you!» That would be my typical reaction to such an event. But if it was evident from the beginning that the interaction that I am having is with a bot, I am more flexible. When the bot makes a mistake, in a sense it makes me feel that still us as humans are so much better. It in a way is a compliment to my intelligence. But when bots fail we should design the way they fail so that it is meaningful.


When a bot fails, forcing the human interacting with it to continue to interact with it is dumb. After many time asking the bot something and he won’t deliver I start to be pissed of. And this even if I knew from the outset that it was a bot. When bot fails, we should allow people to switch to human interaction. It’s like when you are in a shop and ask a vendor something and he doesn’t know the answer. He will tell you: «Sorry I don’t know, I can bring in my colleague who knows more about this topic».