After working on AI-powered applications for a while now, we applied a lot of the lessons learned in building our own bot. Johnny, our DevOps chatbot, enables the release manager skip the confusing discussions with the tekkies and master the release process. It’s using NLP, so you don’t feel like talking to a bot, rather to a teammate. In this session you will learn about the building blocks of Johnny, dive into how each plays its part (pointing out reasoning behind our design decisions), and then we’ll get to do a live demo to see how all fits together. A bot framework is used to pass on information to “the brain”, which decodes your message using Natural Language Processing, and if it’s below a configurable certitude threshold comes and asks you for confirmation/clarification. Once the task to be executed is understood, the “brain” goes on to execute it and hand you over the results.


Comments are closed.

AI is an inherently interesting subject, but Georgiana excelled at presenting a practical guide to setting up a natural-language-learning chatbot, using readily-available components. Well paced, with clear explanation of the components involved and the options for each and examples using IBM Watson Assistant. No unnecessary waffle, and welcoming of questions. Highly recommended.

John Kleiser at 19:28 on 20 Feb 2020

Great, thanks.

Tom Cameron at 08:02 on 21 Feb 2020

Great overview - good technical details too.

Thank you all for joining and for the kind words of appreciation. I’d love to hear about the cool bots you built!