A passage is a "turn" in the chatbot conversation: the chatbot says or does something, then the user responds. Use the Passage Edit screen to create the content your chatbot will send, and determine how the user should respond.
Each passage consists of:
- A Name: a short title that describes the passage.
- The Bot Message: content the chatbot sends to the user, including scripts that might be run.
- The Expected User Response: how the user is expected to respond.
This section can also run a script after the user has responded, or hand control of the chatbot over to a human.
- Inbound Questions: typed responses from the user that should be answered by this passage.
- Outbound Questions: typed responses from the user that should be answered differently in the context of this passage.
You can see whether the passage is in draft, under review, ready for publishing or published by the icon next to the passage name, as well as who last edited the passage, and when. If there are any review notes, you can view and edit them by clicking the comment icon.
Edit the passage name
Change the name of the passage by clicking the edit button next to the name.
Try to keep passage names under 24 characters so they can be easily read in the conversation canvas. If the passage name is longer than 24 characters, you can see the full passage name in the canvas by hovering over the passage card.
Retain the previous conversation context
When the chatbot starts a passage, either from matching a question or following a button or script, the chatbot's conversation context changes to the passage's conversation. This ensures that any questions the user asks are checked against the most recent context, so the conversation feels more natural.
But sometimes you want to use a passage from another conversation without switching the conversation context. For example, if a user says "thanks" in response to something your chatbot does, you want to say "you're welcome"and return seemlessly to the topic at hand, even if the "you're welcome" passage is located in a different conversation.
The Retain the previous conversation context option tells the chatbot to not switch the conversation context when it starts this passage. So if the user says "thanks" in the middle of the Loan Application conversation flow, the chatbot answers with the "you're welcome" passage from the FAQ conversation, but doesn't switch the conversation context to the FAQ. The chatbot then resumes the Loan Application conversation flow as if it had never left, creating a much more natural conversation.
Retain previous conversation vs Resume Conversation
The Resume Previous Conversation Quick Reply feature provides a similar function of returning the user to a previous conversation, with two differences.
Resume Quick Reply
The conversation context will change when the chatbot sends the "you're welcome" passage, which means any questions the user asks after the passage will be matched against the conversation that the "you're welcome" passage is in.
The user must also explicitly ask to return to the previous conversation by pressing the Resume Quick Reply. This can feel unnatural for things like a "you're welcome" response, but may provide a better experience for a user clarifying something.
Retain previous conversation context
The conversation context doesn't change, so any questions the user sends after the "you're welcome" passage are matched against the original conversation. The user is also returned to the previous conversation without any interaction on their part.
Edit the Bot Message content
Add content to your Bot Message by clicking the content type buttons. You can re-order the content by dragging the handles, and delete content using the delete icon.
Some content types need configuration, or can perform additional actions. See Bot Message content types.
Edit the Expected User Response
Select the way you expect a user to respond: by clicking a button or quick reply, typing a question or sending information like an email address or phone number. You can also choose to run a script after the user responds, or pass control of the chatbot to a human.
Defining how the user is expected to respond allows the chatbot to guide the user in using the chatbot correctly (such as pressing buttons or quick replies) when the user gives an unexpected response. You can configure how the chatbot guides the user using the Unexpected Text Response setting in the chatbot's Experience settings.
Some response types need configuration, or can perform additional actions. See Expected User Response types.
Edit the Inbound Questions
Inbound questions are typed user responses that are answered by this passage. For example, if the passage gave instructions on donating to the organisation, Inbound questions might include "how do I donate", or "I want to donate".
To edit inbound questions, click the Inbound Questions card.
For more information about Inbound questions, see Questions.
Edit the Outbound Questions
Outbound questions are responses a user might type that should be answered differently if the user sends them immediately after this passage.
For example, the question "start application" usually takes the user to a passage that asks them what kind of application they want to make. If this passage is about car loan applications, an outbound question of "start application" could take the user directly to the first passage of the car loan application instead.
To edit outbound questions, click the Outbound Questions card.
For more information about Outbound questions, see Questions.
Add a review note
You can add a note for the passage that appears in the Versions page when your team review your chatbot content. Review notes are shared across your whole team, and can be edited through the entire review process. Passages and their linked questions share the same note.
To see or edit the review note, click the comment icon at the top of the passage.
Change the review state
All passages have a review state, which tracks whether the passage has been changed from the published version, and whether those changes have been reviewed. Passages can be in draft, ready to review, ready to publish and published.
When you save changes to a passage that is in draft or ready to review, the passage stays in the same review state it was before unless you choose another state. When you make changes to a passage that is published or ready to publish, the passage is automatically moved to draft.
You can see a passage's review state by the icon next to the passage name:
|Draft. Changes can be made and saved to draft, ready to review and ready to publish.|
|Ready to review. Changes can be made and saved to draft, ready to review and ready to publish.|
|Ready to publish. Making changes will move the passage to draft.|
|Published. Making changes will move the passage to draft.|
Sometimes you'll want to send the passage to a different review state. For example: if you've decided a passage in review needs major changes you might want to save it back to draft, or you've addressed the typo your reviewer picked up and you want to make the passage ready to publish.
You can change the passage's review state when you save it. The review states you can select depend on the user role you have within your team: if you can set changes as ready to publish in the Version page, you can save a passage as ready to publish.
You can see what state you're about to save the passage in by checking the icon next to the save button.
Why can't I set it to published?
Changes to your chatbot can only be published by creating a new version of the chatbot. You can't set a passage status to publish directly. For more information on publishing chatbot content, see Versions.
To change the review state of a passage:
- Click the arrow next to the Save button and select the review status.
- Click Save.
I can't see the ready to publish option
Your user role in your team determines whether you can review and publish chatbot content. If you don't have permission to review drafts or publish content, you won't see those options. See Team and user roles for more details.
See Versions for more information on reviewing chatbot content.