Overview

When your bot does not know how to respond, it will select a fallback response to help your contacts figure out what to do next. For more customization, there is a new page dedicated to editing and managing these responses.

Fallback responses are provided when your bot cannot confidently understand an inbound question.  This can happen when it can't match to anything in our Knowledge Base, or when an understanding is missing an approved answer. When your bot cannot figure out how to respond, it will check for these scenarios to decide which type of fallback response to send:

• Was this an exceptionally long message?
  Your bot will choose randomly from the Long Question list.
  • If a contact’s message is longer than 120 characters and the bot cannot understand the question, it will suggest shorter responses.
  • This is also an indicator that someone needs additional support from a human who can take a complex or specific personal issue and is a great time to suggest #followup or another way contact the school for help. You can set up the #followup command by going to the Commands section of your Knowledge Base  
  • Check out #Commands Page for more information on how to set it up
• Is it just a question your bot doesn’t know yet?
  Your bot will choose randomly from the General list for all other questions.
  • For all other scenarios, your bot will choose randomly from this list
  • A variety of responses make your bot seem smarter and allow for personality.
  • These can also include helpful tips and suggestions on occasion to refresh your contacts on what to do next.
• Is this a new contact?
  Your bot will send the next response in the Training series 
  • New contacts need extra informative fallback responses with details about what happens when they confuse the bot.
  • This series is intended to set realistic expectations for how accurate the bot can be, what topics it’s most familiar with, or steps they can take to get a better answer.
  • Note that contacts will receive each of these, in order, before receiving any General Fallbacks. However, if a contact receives a single personalized Training series Fallback (see below), then that completes the Training series for that contact.

To learn more about filling gaps within your Knowledge Base and reducing avoidable fallback responses, please see Knowledge Maintenance & Webchat Feedback. 

Check out Mainstay’s writing tips for Fallback Responses: How to write a Fallback Response

Viewing Fallback Responses

1. Log in to the Mainstay platform and navigate to the Knowledge Base.
2. In the left sidebar panel, click the section called Fallback Responses.

Click anywhere in one of the rows to see the Fallback Responses sent in that scenario. A panel will open from the right with all the responses for that particular situation.

Editing Fallback Responses

1. Click anywhere in one of the rows to see the Fallback Responses sent in that scenario.
2. A panel will open from the right with all the responses currently used by your bot for this scenario.
3. Click the response you want to edit.
4. Use the text editor to rewrite or personalize a response.

Adding Fallback Responses

1. Click anywhere in one of the rows to see the Fallback Responses sent in that scenario.
2. A panel will open from the right with all the responses currently used by your bot for this scenario.
3. At the bottom of this list, click + Add Response
4. Use the text editor to write a new response.
5. When you are finished, click save.

Personalized Fallback Responses

Like normal understandings, you can provide personalized responses for fallbacks.

1. In the Fallback Responses panel, scroll down to Personalized Responses.
2. Click + Add Response to add a new personalized response, or click into an existing one to edit it
3. Personalized Fallback Responses can include embedded Contact Fields and/or be targeted to a particular Audience. At least one of these features must be in use to save a personalized fallback response.
4. As with Personalized Responses for normal understandings, it's possible a contact will be eligible to receive more than one. In that case, they will receive the first one in the list that they are eligible for. You can use the drag-and-drop handlers at the top of each response to determine the priority order.

Using Mainstay Suggestions

1. Your account is pre-loaded with all the responses we recommend as well as any custom responses you added with the help of your Success Manager before the release of this feature.
2. To see all of our suggested responses, click anywhere in one of the rows to see the Fallback Responses sent in that scenario.
3. At the very bottom of this list, click the gray box that says “Not sure what to write?”
4. Click + Add on any of the responses you like, and they will be added to your account.
5. To add every suggested response, click + Add all to my account
6. All suggested responses are editable after they have been added.

Deleting a Fallback Response

1. Click anywhere in one of the rows to see the Fallback Responses sent in that scenario.
2. A panel will open from the right with all the responses currently used by your bot for this scenario.
3. Hover on the response you want to delete.
4. Click the trash can in the top right corner of the response. Note that you cannot delete the last remaining non-personalized fallback response for that type.