We have an issue wherein our bot triggers the “Lists all actions and FAQs for a bot” whenever the word “instruction” is inputted by the user. We were not expecting this standard message since we have configured a dialog task to run for the Intent not Identified event
For other intents that are not identified, this dialog task is successfully called. Please see these screenshots below of coffee and instructions utterance which should have both triggered the task for Intent not identified event.
The standard response here is for “help”, which is fallback processing (i.e. when nothing else matches) in the platform that will respond with a list of tasks. “instruction” is one of the words that triggers this fallback (along with help, guide, guidance, assist, manual).
The standard response message can be obviously be changed.
If a real task has training for these words then that would be executed instead.
You can also add a concept named ~disable_help with some word (doesn’t matter what it is) that will disable this help standard response. But it would be up to you as a bot developer to have some sensible response when a user asks for help.