Agentforce Platform | Extract Fields and Values from User Input
Extracts fields and their corresponding values from a record based on the user's
request. The output of this action is used as the input for the Update Record action. For example,
when a user wants to edit a field on a record, the Extract Fields and Values from User Input
action is triggered first to retrieve the field and its value, and then the Update Record action
saves the updated extracted value to the record.
Required Editions
Available in: Lightning Experience
Available in: Enterprise, Performance, Unlimited, and
Developer Editions with Foundations, or Agentforce 1 or Einstein
1 Editions
Does this action execute one or more prompt templates?
Yes
Considerations
Only existing records that are searchable and User Interface API enabled can be updated by
this action.
This action only supports objects with layouts.
The action only returns fields from the record’s page layout that can be updated or created.
Read-only fields are filtered out even if requested by the user.
Users must have permission to view and edit the record and fields they want to update. Users
can update one record at a time.
This action doesn’t support record creation. For example, users can’t ask the AI agent to
log a call because that requires creating a record.
This action only supports the editing of records by updating record fields and their
corresponding values. Other types of record changes—such as cloning, merging, associating a
related record, transferring account ownership, or submitting a quote—aren’t supported.
This action can’t create a new field on an object’s layout. For example, if a user enters,
“Update the service start date to 12/10/2024,” the request will fail if the Service Start Date
field doesn’t already exist.
If a user wants to edit a multi-select picklist, only one value is supported per
update.
In percent fields, there's no calculation of a percentage. The action assumes that the value
the user entered is already a percentage value.
Conversion calculations in currency fields aren’t supported. For example, if a field stores
an amount in dollars and the user asks to update the amount to 500 euros, the value will be
saved as 500 dollars.
In numeric fields, only the shorthand notation for thousands (K), millions (M), and billions
(B) is supported. For example, "2K" will be interpreted as “2000.” However, the values of
numeric fields can’t be converted. For example, if a field is named Number of Years and the
user tries to update the value to "36 months,” the value will be changed to "36 years" rather
than "3 years."
Updates from a related record aren't supported for many-to-many relationships. For example,
users can’t make a request such as: “Remove the contact John Smith from the Acme
opportunity.”
Users can’t append a value to an existing value in a field. The value the user provides
overwrites any existing value.
We use three kinds of cookies on our websites: required, functional, and advertising. You can choose whether functional and advertising cookies apply. Click on the different cookie categories to find out more about each category and to change the default settings.
Privacy Statement
Required Cookies
Always Active
Required cookies are necessary for basic website functionality. Some examples include: session cookies needed to transmit the website, authentication cookies, and security cookies.
Functional Cookies
Functional cookies enhance functions, performance, and services on the website. Some examples include: cookies used to analyze site traffic, cookies used for market research, and cookies used to display advertising that is not directed to a particular individual.
Advertising Cookies
Advertising cookies track activity across websites in order to understand a viewer’s interests, and direct them specific marketing. Some examples include: cookies used for remarketing, or interest-based advertising.