You are here:
Generalist Agent (Beta) in Agentforce Operations
Use the Generalist Agent (beta) to automate a broad range of tasks in Agentforce Operations. Supported tasks include routing decisions, calculations, data source queries, table operations, and document processing.
Required Editions
| License Required |
|---|
This feature is a workspace that’s external to your Salesforce org and requires an Agentforce Operations license. To purchase an Agentforce Operations license, contact your Salesforce account executive. |
When to Use the Generalist Agent
The Generalist Agent uses Python to complete tasks, which makes it more flexible and accurate for complex logic, large datasets, and multi-step operations. It covers the capabilities of the Flash, Regrello, Tabular, and Document agents.
For document extraction tasks, use the Document Reader Agent. To process tabular data in CSV files by using Excel, use the Excel Agent.
Capabilities
The Generalist Agent supports a broad range of capabilities across logic and reasoning, document processing, table operations, and calculations.
| Category | Capability | Description | Example |
|---|---|---|---|
| General logic and reasoning | Logic-based routing | Deterministically evaluate logic conditions. | If an invoice total is more than $50,000, require the finance team to review it before paying. |
| Rule-based decision making | Triage or route based on configured conditions. | Identify which team to route an email to: Support for product issues, Finance for invoices, or Legal for contracts. | |
| Document classification | Triage or route tasks by evaluating document or image files by using simple logic. | Classify the type of damage shown in an image of a package: water damage, shortage, physical damage, or no damage. | |
| Task approval | Approve tasks based on conditions in the task description. | Approve only if invoice price equals contract price. | |
| Document analysis and generation | Document summary | Summarize one or more documents. | Summarize the intellectual property clause of the provided contract. |
| Word template-based document generation | Generate a .docx file from a Word template by replacing text placeholders with task
field values. Text placeholders in the .docx template use the format {{ field_name
}}. |
Generate a contract document from a template by using a customer's name, the contract's start date, and the contract's end date. | |
| Word document comparison | Compare two .docx files and produce a table of all differences, including comments. | List all differences between the originally proposed contract and the latest proposal returned from a customer. | |
| PDF generation | Generate a PDF by formatting task field values and documents into structured output. | Format a supplier form, scores, and scoring rationale into a report with a summary at the top. | |
| PDF page extraction | Extract pages from a PDF into a new PDF based on criteria in the task description. | Extract all pages in a contract that contain intellectual property terms into a new PDF. | |
| Table operations | Table filtering | Create a table containing only the rows from the original table that meet conditions stated in the task description. | Create a table containing all rows from a comprehensive parts list where the country of origin isn’t the USA. |
| Calculated column generation | Copy an existing table into a new CSV file and add one or more columns with values calculated from existing row data. | Copy a purchase order table into a new CSV file and add a Total
column that contains each row's Quantity multiplied by its
UnitPrice. |
|
| Column total calculation | Calculate the sum of values in a numeric column of an existing table. Store the result in a Request information number or currency field. | Calculate the sum of the Total column in a line items table and store
the result in the Subtotal currency field. |
|
| Table merging | Merge two or more tables that each have a column with an identical header and compatible data types. Store the combined data as a new CSV file. Optionally, store rows without a match in a separate CSV file. | Join the Item List CSV and the Tariff Rates CSV by using the hs_code
column. Store the combined data in Combined Data CSV. Store any unmatched rows from Item
List CSV in Item Exceptions CSV. |
|
| Table comparison | Compare two versions of a table that each have at least one column with an identical header and compatible data. Store rows where values differ in a new CSV file. Store rows that exist in only one version in the same or a separate CSV file. | Compare the Original Pricing CSV and the Updated Pricing CSV by using the
ItemNumber column. Store rows where ItemPrice values
differ, and rows from Updated Pricing CSV with no matching ItemNumber in
Original Pricing CSV, in Pricing Differences CSV. |
|
| Data aggregation | Group rows by a shared value. Store one summary row per group with calculated values in a new CSV file. | Group rows by Category and store the TotalAmount for
each category in a new CSV file. |
|
| Data source SQL operations | Query rows from a data source by using SQL operations and store the results in a new CSV file. | Find all suppliers where Country is USA and
NAICSCode is 54151 and store the results in a new CSV file. |
|
| Data source natural language search | Search a data source for rows that match the meaning of a phrase and store the results in a new CSV file. | Search Customer Claims / Incident Notes for claims related to shipment damage and store the results in Damage Claims CSV. | |
| Table format conversion | Convert an existing CSV or .xlsx file to the other format. You can also output tabular data generated during the task as a CSV or .xlsx file. | Store the updated Tariff Rate table as an .xlsx file in Tariff Rate XLSX. | |
| Other | Numeric calculations | Perform calculations by using standard arithmetic operators: +, -, *, /. Use parentheses to control the order of operations. | Calculate the landed cost by using the formula: (ProductCost +
Freight + Duties + Insurance +
HandlingFees) * TariffRate. |
| Date calculations | Find the number of days between two dates, or add or subtract days from a date. | Calculate the contract end date by adding 365 days to
ContractStartDate and store the result in
ContractEndDate. |
|
| Text operations | Perform text operations on field values. Operations include combining strings, extracting substrings, counting characters, and converting numbers or dates to text. | Combine SupplierName and ContractNumber to create a
document title in the format SupplierName
_ContractNumber. |
Limitations
- The maximum size for an individual document in an agent task is 500 MB.
- The maximum combined size of all documents in a single agent task is 1 GB.
- If the task documents exceed either limit, the agent task fails with an error message.

