Conversational AI is fast catching up and it will not be one of the multi-channel option rather a predominant channel to reach the target audience, this is not a product of wishful thinking rather a prediction punctuated by facts, here are top 4 reasons:
a) Ease of use — I can speak with my Alexa/Siri/Google assistant even from a shower or while playing my kids
b) Transcending the barrier of literacy it penetrates the fortified walls of the baby boomers
c) Most Loyal Pal — I had never seen my digital device, getting infuriated with my repeated questions, rather it makes all possible attempts to understand me better , this endless patience is one feature which will always differentiate Bot from an human (wish spouses can be like this !!)
d) Cost- Needless to say we get 24/7 coverage at no minimal hourly billing, but among all other benefits this is a given and is available in mobile app as well
Smitten by such prospects, when a naive developer starts downloading Anaconda/Python Distribution to start with a chat-bot they get dissuaded by enormity of Environment Setting(predominantly in windows).
Trending Chatbot Tutorials
This blog is an attempt to put together road-signs to help navigate weary traveler through this complex jungle
1- 'git' is not recognised as an internal or external command,
operable program or batch file.
Solution : Download Git for windows , installation alone doesn't solve the problem, you need to update the environment variable -PATH in windows setting. Will this suffice? no, if Terminal is open in your Anaconda then you need to close and open it again for this change to be affected
2- If you had not installed VC++ build tools from Visual Studio , many of RASA packages will not be resolved accurately leading to errors such as ModuleNotFoundError, etc
3- If you are linking Spacy you need to do it with admin mode, logging in as admin will not be sufficient rather command prompt should be executed in admin mode(If you are using anaconda navigator, right click and select “Run as administrator” option”) .Most misleading part is at times without admin access message appears as “Linking is successful” whereas just above this status you could see “You do not have sufficient privilege to perform this operation
4-Minimal hardware requirements : With Pentium dual core processors, Tensor flow wheel will not be fully installed rather will throw an error ImportError: DLL load failed with error code -1073741795 , even with CPU option alone, so for the CPU’s which doesn't have the AVX instructions you can follow this thread to download appropriate wheels
5- Other constraint with RASA framework is its rapidly changing versions , some of the features gets deprecated with no forward compatibility leading to complete rewrite, you can refer to my recent Github link for a working version of WeatherBot
With NPM GUI for drafting the intents and a super cool Rasa Interactive trainer, RASA is one of the best and simple framework available for building on-prem chatbots, for the starters this videos may help. Good luck on your BOT journey !!
Don’t forget to give us your 👏 !
Common Mistakes in RASA ChatBot Environment setup (Design Part 2 of 3) was originally published in Chatbots Life on Medium, where people are continuing the conversation by highlighting and responding to this story.