Updates in Knowledge Graph | Yext Hitchhikers Platform
What You’ll Learn
In this section, you will learn:
- An overview of the new Knowledge Graph features from the Spring ‘23 Release
Computed Field Values (Preview)
Computed Field Values (aka Computed Values) allows businesses to configure an algorithm to automatically populate the value for a particular field in the Knowledge Graph.
In order for a field to have its value computed, a user must define a Computation Method for the field in question. Computation Methods are defined on a field level, as computation logic will depend on the field and the content you want to store in that field.
Defining a Computation Method
Computation Methods are defined on the field-level, as computation logic will depend on the field and the content you want to store in that field.
By default, a field in the Knowledge Graph will not have a computation method assigned to it, as the field’s value is usually populated explicitly for each entity. Selecting and defining a Computation Method for a field enables the value for that field to be computed automatically by the Knowledge Graph system.
Content Generation and Computed Field Values
Within Computed Field Values, a class of Computation Methods leverage Content Generation. These methods use generative AI in the algorithm to compute a field’s value.
Today, the only available computation methods for Computed Field Values fall within that Content Generation group. However, in the coming releases, we plan to expand the list of computation methods offered significantly, to include additional AI-based Computation Methods, as well as non-AI methods.
Because of this, Computed Field Values and Content Generation should not be thought of as 1:1. Computed Field Values is a Knowledge Graph Feature. Content Generation is a class of infrastructure which will be deployed across the Yext Product Suite, beginning with applications in Computed Field Values.
How to Set Up Computed Values
To define a computation method:
- Navigate to Knowledge Graph > Configuration.
- Click on the Fields tile.
- Click into an existing custom field (or create a new custom field).
- Click on the Computation Method tab.
- Select the desired Computation Method from the drop-down.
- Add the relevant inputs.
The inputs are a set of parameters (think: variables or instructions), that are received when a computation is triggered, in order to execute the computation. These inputs will vary depending on the selected Computation Method. For example, if you select the “Rewrite Text” method (as seen above), you will need to enter the text you would like to have rewritten, the desired output format, and any additional prompt instructions. You can supply embedded fields as the value for inputs, which allows the system to dynamically grab the value for the supplied field to “resolve” the input when running a computation.
Prompt Writing for Content Generation based Computation Methods
When leveraging Content Generation based Computation Methods, you should be thinking about the fact that you are writing prompts for a generative AI model. In order to get the desired results, you need to ensure that the information you want the model to consider is included in a computation when configuring inputs. In each Content Generation computation method, you can add Additional Instructions in prose, in order to account for any nuances. To learn more about how to write effective prompts, visit our .
Triggering a Computation
Once a Computation Method is configured for a field, the actual computation must be executed for a particular Entity Profile. The act of invoking a given computation is referred to as triggering the computation.
Today, the only way to trigger a computation is through Entity Edit, for a single Entity Profile at a time.
To trigger a computation:
- Navigate to the desired entity.
- Click on a field that has a Computation Method defined for it.
- Click on the Wand icon displayed in the bottom right corner of the field to trigger the computation method to be executed.
- The Wand icon will only display for fields that have a Computation Method defined for it.
- Review the computed value, make any changes if you’d like, then click Save.
How This impacts You
Computed Field Values allows users to populate Knowledge Graph fields dynamically, based on specific algorithms and inputs you configure.
Today you can start using Computed Field Values for computations powered by Content Generation — allowing you to generate net-new content and parse existing content. Such as:
- Write a long-form business description based on structured data stored about that business.
- Generate a meta description for a webpage, using existing page content.
- Write a bio for your professionals, based on an individual’s attributes such as experience or speciality.
In future releases we plan to expand to a variety of new methods and functionality — increasing the impact of Computed Field Values in the Knowledge Graph to include things like:
- Automatic computations, triggered by input field updates
- Bulk computations, to compute field values for multiple entities at once
- Suggestions workflows, to ensure human review before a computed value is applied to an entity
To turn this feature on in your account, navigate to Account Settings > Account Features and select Preview Feature: Computed Field Values.
The use of Content Generation in Yext requires the purchase of a Content Generation subscription. Please reach out to your account team to learn about Content Generation pricing, or the option for free capacity.
File Field Type
We’ve added a new File field type to the Knowledge Graph, allowing users to upload and host any type of file (image, video, PDF, etc.) in Yext. Once a field of type file is configured and enabled on an entity type, files can be uploaded into the field to be made available at a Yext-hosted URL. Files can then be served in Pages, Search, or other custom headless experiences — in a future release, the content of Files hosted in the Knowledge Graph will be searchable by our algorithms for Search and Chat experiences.
To leverage this feature you will need to create a custom field with the new File type. Once you do this you will be able to add the relevant file (either by uploading a file, or entering a URL) and then leverage that as needed in any of your downstream systems.
How This impacts You
This new File field type ensures that users can use the Knowledge Graph as a comprehensive store of their brand content, hosted by Yext and primed for delivery to any digital experience.
To turn this feature on in your account, navigate to Account Settings > Account Features and select Spring ‘23: File Field Type (early access).
New Rich Text & Markdown Field Types (Preview)
Rich-Text v2 Field Type
The new Rich-Text v2 field type is a significant improvement over our existing Rich-Text Field Type, as it supports new features such as:
- Text colors
- Text highlighting
This is a new field type that is distinct from the existing Rich-Text field. To leverage this field you will need to add a new Custom Field of the Rich-Text v2 Field Type. Over time we plan to replace the existing Rich-Text Field Type with this new Field Type.
Markdown Field Type
This field type includes capabilities such as a formatting toolbar to insert the appropriate Markdown syntax directly in the text editor, and a side-by-side rendered view of the content. Users can also include HTML directly in the field as markup for any formatting that is not specifically supported in Markdown syntax.
To leverage this field you will need to add a new Custom Field of the Markdown Field Type, and make that available to the relevant entities.
How This Impacts You — Rich-Text v2 Field Type
Benefits and Intended Use Case
This field is intended for any content author in Yext, and will be the future of how people write richly formatted content in the Knowledge Graph. This field is very flexible and easily to use, allowing anyone to create content with the formatting they need.
The Rich-Text v2 field type is primarily intended for authoring content in Knowledge Graph. If you do not intend to author directly in Knowledge Graph, but wish to pass formatted content through Knowledge Graph to be served on your digital experiences, we recommend leveraging the Markdown field type.
In order to render content from the Rich-Text v2 field type on a website, the content needs to eventually be converted into HTML.
At this time, we have support for a React component in our React library that will convert content in this Rich-Text field to HTML. However, tooling is only supported for React-based front-end experiences. At this time, we do not recommend attempting to render this field type on non-React front-ends.
In the future we plan to support conversion in the platform.
How This impacts You — Markdown Field Type
Benefits and Intended Use Case
This new field type will ensure that technical users can author Markdown and HTML formatted content directly in the Knowledge Graph with the utmost flexibility.
This new Markdown field type also makes it very easy to ingest content from external sources, as the Markdown or HTML formatted content will be preserved and stored as they are, and there is no conversion that needs to take place.
In order to render content from the Markdown field type on a website, the content needs to eventually be converted into HTML. Since Github Flavored Markdown is a standard format, there are a variety of well maintained libraries that handle converting GFM to HTML which can be leveraged to achieve this goal.
In a coming release, we plan to support conversion from Markdown to HTML in Streams (and therefore Pages, Search, and the Content API), as well as in the Entities API (Management API). Once this is available, developers can simply request HTML instead of converting the content themselves.
To turn this feature on in your account, navigate to Account Settings > Account Features and select Preview Feature: Markdown and Lexical Rich-Text Field Types.
Entity History UI
This release we have introduced a dedicated Entity History UI within the platform, providing users with a comprehensive view of all changes made to the Knowledge Graph.
In your account, there are three ways to access this new screen:
- Via Knowledge Graph > Configuration > Entity History.
- In the top right-hand corner of Entity Search
- In the top right-hand corner of Entity edit (and this path will give you a view that is automatically filtered to that entity)
Each row in the Entity History table corresponds to an update on an entity based on a variety of event types — such as Entity Created, Field Updated, or Entity Deleted. For details on the specific event types you can view, visit our Entity History reference documentation.
Clicking View next to a corresponding row will open a modal with the details of the change. For example, for a field update you will be able to see the updated content, and the previous value to view a log of the changes. From there, you can also restore previous values if an update is incorrect, or was made in error.
How This impacts You
This update is a further enhancement to our audit and content management capabilities. With this improvement, users can easily investigate historical information about updates and reconcile any unwanted updates across the Knowledge Graph directly from the Yext platform with the click of a few buttons.
To turn this feature on in your account, navigate to Account Settings > Account Features and select Spring ‘23: Knowledge Graph Entity History (early access).