Understandings are collections of related question and answer pairings that your bot can use to answer incoming questions from learners. Each understanding will have a main Summary question and subsequent synonym questions that try to capture different variations of ways to ask the Summary question.
Each understanding in the Knowledge Base contains various synonyms that expand the AI's ability to properly match incoming questions to existing content in the Knowledge Base. An understanding is made up of the following components:
Questions | Responses |
The Summary Question represents the intent or central meaning of an understanding. It should be unique and simple. | A Default Response is the general answer to the understanding's questions. |
A Personalized Response is a more specific response you can customize based on specific criteria if the answers to your questions can vary depending on the scenario. |
|
|
An Interactive Response is a response that utilizes a script to clarify the intent of a question, and connect the learner with important contextual information |
Other tools to leverage in Responses: |
Tips
Making understandings
- Understandings should be as unique as possible. Overlap in questions may cause unintended bot mismatching to occur.
- We typically recommend that each understanding have at least 10-20 additional synonymous questions to help the bot effectively match to understandings.
- Questions on an understanding should be representative of simple ways that someone could ask a question. Long, multi-sentence questions should be avoided as well as overlapping content with other understandings/questions.
-
- The maximum length for a new questions should be around 140-160 characters
-
- After making understandings or adding synonym questions, you can always use the Test the Bot feature to double check your work and verify that your updates are working as you intended.
- For a more detailed guide, see Creating a New Understanding
Crafting responses
- Responses should not begin with a "yes" or "no". Learners can ask questions yes or no questions in different ways, and answers should focus on the information that is most important to learners.
- Example: Questions around campus safety
All three of these questions have the same intention and could live on the same understanding. preceding the answers to this understanding with a yes or no could confuse the contact or make it unclear if their question is answered.
-
- "is there a campus security office"
- "is there not a campus security office"
- "can you tell me about the campus security office"
-
- Example: Questions around campus safety
Matching
- Understandings need to be Approved in order to provide a response. Understandings that are Not Approved will provide fallback responses. For understandings with multiple approved responses, the AI will select any approved answer to give back to the learner.
- Our AI analyzes and interprets each question at a time, and does not factor in previous conversational context when deciding which understanding's question is the closest match.
Summary Question
The summary question represents the intent or main purpose of an understanding. It is a simple and clear representation of all other question synonyms. It is not editable after an understanding has been created. For example, you can see the summary question, "how do I clear an emergency contact info hold," in the table view in the Knowledge Base:
The Summary Question also appears at the top of the Understanding edit panel, above your bot's answers, and at the top of the Questions section, above Organization Questions and Global Questions.
Questions
Collectively, all question synonyms within an understanding are simply called Questions. These questions are synonymous or parallel in meaning to the summary question and to one another. A bot’s questions include two categories: Organization Questions and Questions. Both types of questions help sharpen a bot’s AI, and enable it to better match incoming questions to an understanding covering the same subject.
Organization Questions
Organization Questions can be used to tailor any understanding more closely to your institution. These questions are specific to each bot’s Knowledge Base and can be created, edited, and deleted in the platform. To be most beneficial to a bot, Organization Questions should closely match the intent or subject of the Summary Question of an understanding.
Adding Organization Questions is entirely optional. It's a great way to customize your bot's Knowledge Base and build off of the Global Question Collection, but it's by no means required for the AI.
To add an Organization Question:
- Click the + Add Question button
- Type in the new question synonym
- When you are ready to save, click Create Question or use the return key on your keyboard
- If you do not want to save your new question, click Cancel
Any Organization Question you create can be edited later on by clicking the question or deleted using the trash can icon to the right of the question.
Global Questions
Global Questions are a collection of questions that are included in all bots’ Understandings. They are synonyms of the summary question that have been identified and added to an understanding by ’s knowledge-building team. While you cannot edit into these questions, you may deactivate them individually if they do not fit within a bot’s Knowledge Base or do not apply to an institution.
To activate or deactivate any Global Question, click on the toggle next to that question. Once a question has been deactivated, it will appear with a grey toggle in a separate section, below all the questions that are still active.
Global Questions occasionally include words beginning with an “@” symbol, like @college or @orientation, which are generic representations of an institution’s information called entities. These allow Global Questions to be used across different bots’ Knowledge Bases.
Comments
0 comments
Article is closed for comments.