The 5 things we've learned about building bots – Chatbots Magazine

We built BOTbot as an experiment to learn about self-service bot creation platforms, user/bot interaction, bot promotion and management. Below we run down the process we took to creating a code-free MVP, the results from our initial user interactions and how scripted journeys, rails, NLP and AI all have their place in business bots -- dependent on use case. After looking at various'quick&easy' bot platforms we decided to use Chatfuel because: We continue to be impressed with Chatfuel's simplicity and use it for demos and tests. But we don't use Chatfuel for live production bots because: Having chosen the platform to build on we launched BotBot without buttons, optimistically hoping to use NLP to create an open conversational dialogue about chatbots. Limited domains, where a data set is filtered across a handful of axis (e.g.