Story Board and User Feedback
Story Board is where utterances, intents, and entities transform into a functional SUVA bot. In the SearchUnify parlance, a story consists of nodes. You can connect several nodes to create a story.
Node vs Intent
Nodes are separate from intents. A way to comprehend nodes is to consider a scenario, where a user has three ways to proceed when they show an interest in "search clients":
Find platforms where a search client be installed,
Find instructions to install a search client on a supported platform, or
Discover the list of configurations available after installation.
Intents are used in the parents and child nodes to trigger the responses from the appropriate node of a story. This facilitates users to cover multiple stories within one single story. Although it is optional to add an intent to a child node.
An advantage of separating nodes from intents is that once an intent has been detected, you can create comprehensive stories around it rather fast. When there are several nodes added to a story, you can use the search bar to locate a node.
Compose a Story
- When you edit a Virtual Agent, go to Story Board. You will see four pre-configured stories - Greeting, GoodBye, Role, Role info
- To create a new story, click Compose Story.
- A dialog with three fields will appear, of which the first two are mandatory.
- Name. The name of your story that you cannot change. It is not visible to instance users. Names consist of only letters and underscores.
- Label. The name of your story that can be changed anytime. It is visible to instance users. Names can contain any Unicode character.
- Description. A short summary of the story.
- Save the story.
- Click on the Start node. Admins can add as many nodes as they want (both as parent and child nodes)
- In the dialog that has opened, enter Node Name and Node Label. After that, pick a trigger in Select Intent. If none of the existing intents matches your story, you can use Create New Intent to write a trigger from scratch.
Note: a) For a newly-created intent, utterances are added from the Intents section. An error is thrown when insufficient utterances are available.
b) Adding an intent is mandatory for parent nodes while it's optional for child nodes. One intent can only be used for one node only and if a node is deleted, the intent assigned with it becomes available to be used again.
- Configure Response. Once you have assigned a Name, Label, and Intent to a story, you can save settings and return anytime to configure responses. Alternatively, you can configure them immediately. Select Response Type offers several response types in four categories:
a) Simple: Text, External Link, and Pause/Delay. For more information, check out: Get to Know SUVA Response Types: Simple .
b) Selection: Options and Prompt. For details, check out Get to Know SUVA Response Types: Selection .Get to Know SUVA Response Types: Adapters.
d) Action: Webhook. For details, check out Get to Know SUVA Response Types: Adapters
Create Snippet using KB Article. Generate snippet responses using the KB articles in a story. For details, check out Create Snippet Using Knowledge Base (KB) Articles
User Feedback (Non-LLM Chatbots)
User Feedback doesn't work on Salesforce legacy search clients.
Don't check Disable User Input. It inhibits User Feedback from functioning.
Setting up a SUVA agent is a start. Immediately after it goes live, you want to ensure that users are getting relevant answers. Digging through chat logs works in the early stages, when users are few, but this approach doesn't scale. Moreover, users never get to know how hard you are working to improve user experience. User Feedback offers a solution to both of these challenges.
User Feedback asks users to rate conversations. The ratings are stored in a report, where you can review those. It's more efficient than perusing through chat logs. From the users' perspective, a simple feedback question assures them that someone is really interested in helping them. They feel valued.
Unlike many other features, User Feedback doesn't have to be configured. It's linked with the first node in each story and is turned on by default. You can turn it off by toggling User Feedback to the left.
On the users' side, a feedback question—Did you find what you were looking for?—is displayed after conversations.
The user can thumbs up () or thumbs down () their experience.
When is clicked, the user gets the message: "Thank you! We are glad that we could help."
On the other hand, when is clicked, the user sees a chatbot with the prompt "Please elaborate your experience." The user can add details or click Skip.
Last updated: Tuesday, December 12, 2023
Or, send us your review at firstname.lastname@example.org