Production Checklist

You are viewing documentation for a legacy Rasa product. If you want to get started building assistants with Rasa please check out our latest documentation here.

User feedback should continue to drive your development when your assistant is in production. Below you will find a list of tasks you should going into production.

Share [

• Try sharing your assistant with at least 50 test users
• Connect and test your production channel with users

Review [

• Decide how frequently you will read through new conversations (e.g. daily)

Annotate [

• Collect 75 NLU training examples per intent per intent at a minimum
• Ensure over 90% of these examples are from real conversations with users

Test [

• Ensure each intent, entity, action, etc. is in your test stories 5+ times
• Set up continuous integration and deployment to ship updates reliably

Track [

• Come up with a way to identify successful or unsuccessful conversations
• Use the Rasa Enterprise API to automatically tag those conversations, if possible

Fix [

• Determine how you will track issues (e.g. Trello) that need to be fixed