Chatbots attacking :)
Sometimes I think that world just goes crazy. Is this just my feeling?
Chatbots – all those articles on LinkedIn and across the google search about the chatbots which are intelligent and will save you a lot of costs – well If you can take a deeper look, you will realize, that those chatbots has mainly one intelligent part – NLP for intent/entity recognition, than maybe there can be some cognitive service connected to the chatbot itself and sometimes (mainly from the vendors other than google or Microsoft) administration global portal for design of all the parts in the chatbot, mainly already mentioned intents, entities, utterances, also for definition of questions & answers and sometimes also conversations tree. But what else? Where is the intelligent part? Does chatbots learn from previous conversations? Mostly not if there isn’t any deep learning mechanism/algorithm for it.
- Are we ready for putting chatbots to the production? Mostly not
- Why? Well we have too much marketing before our eyes, that we are hardly able to think about real use cases where chatbots saves time without need of extra headcount for administration
- What to do? Strategy planning, small to bigger use cases, meta-chatbots / micro services principle of decomposition
I understand that chatbots are extremely useful even todays. Generally. Internally and externally, however one has to be superuser about the concept / use case. Is chatbot used as a frontend for some automation requests and output of the results? Or some general FAQ? This is fine of course. IT saves some time. Imagine bigger implementation with nonstop bigger content for FAQ, static processes etc. This is not already ready as someone still must be there for corrections and other stuff.