Stay up to date with market insights, trends and tutorials about AI, Messaging and chatbots.
Covering the industry around AI for customer service, customer experience automation, chatbots for Messenger, WhatsApp or other Messaging apps and the latest in the NLP developments, such as BERT, ULMFiT or ELMO.
The Head of Product for Poncho bot shares the design decisions they took, and how they’ve learned from the user interactions with the bot. I share his opinion about not trying be too smart. I like keeping “Being Smart by Playing Dumb” in mind during the design process.
Great article from the Head of Slack Developer Relations, pointing out that building a bot is cool and all, but we really need to think if our service really takes advantage of a conversational interface. I loved this quote: “exposing your service with a bot will not make it awesome, unless it is already awesome”.
They thought they were talking to a bot, but it was actually a person. What he learned? People are totally unpredictable. Not surprising. One of our clients have an SMS bot where he states a very simple question asking only for a yes/no. Thousands of different responses, with examples like “well, hopefully tomorrow I can say a yes but now I am not sure”. Take that and process it with wit.ai, yep.
It shows the state of bots from the perspective of a natural conversation with its users. I agree we are far from it, and that there has been way too much hype about this topic already. What surprised me about this article are some quotes from people working directly with chat bots that seem to hate the chat bot space.
This post got lots of attention this week. Explains the possibilities of having a self-bot that others can interact with. The bot that comes to mind for this concept could be the next generation of x.ai, where the assistant knows my personality well and fully impersonates me to schedule a meeting. That would blow my mind.