The Salesforce action pack contains the following actions:

  • Search Salesforce with SOQL
  • Update Salesforce Opportunity

​
Overview

Glean’s Salesforce actions let teams efficiently query and update Salesforce data within Glean agents. Users can run real-time SOQL queries and native searches to retrieve insights across both standard and custom objects, supporting lookups and aggregations for GTM and finance use cases. Additionally, agents can automate updates to Salesforce Opportunities—using AI-powered field suggestions and streamlined user reviews—to keep deal information current and minimize manual data entry. These combined capabilities help maintain accurate pipelines while reducing effort for sales teams.

​
Salesforce Search with SOQL

​
Key Features

  • SOQL Integration: Lets users query Salesforce using Salesforce Object Query Language for targeted, accurate answers.
  • Real-Time Data Retrieval: Fetches up-to-date records from standard (Accounts, Opportunities, Leads, Contacts, Cases) and custom objects (if configured).
  • Customizable Data Model: Admins can describe their Salesforce schema to optimize query relevance.
  • Multi-Record and Aggregated Queries: Supports filtered, basic computed, and aggregated queries (requires schema hints).
  • Single-Record and Field Lookup: Enables retrieval of information for individual Salesforce records, including those with custom fields.

​
Usage Instructions

​
Set Up in Agent Builder

  1. In the Agent Builder, select the Search Salesforce with SOQL action by navigating to Select Step > Actions > By datasource > Salesforce Actions > Search Salesforce with SOQL.
  2. Use and test the action with queries such as SELECT * FROM Opportunity WHERE Amount > 100000. If you need help building the SOQL, use the SOQL builder agent linked in the step instructions. Refer to the Best Practices section for tips on getting the best results from your SOQL builder agent.
  3. For advanced usage, add an Analyze Data step.

​
Examples

  • Opportunity Status: “What is the stage of the Acme opportunity?”
  • Custom Object Lookup: “Who is the CSM for the Acme account?”
  • Aggregation: “Total ARR of all the opportunities closed from June to August.”
  • Use Case Association: “What are the use cases associated with the Acme account?”

​
Best Practices and Troubleshooting

  • Initial admin setup for the Salesforce data model is required in the Connectors setup page.

​
Known Limitations

  • Out-of-the-box support for standard objects; custom objects require additional configuration.
  • Cannot search multiple records at once; queries must be split for multi-entity lookups.
  • Search results exceeding 1000 rows will not be shown.
  • Referential (Service Cloud) objects such as Knowledge Bases and Cases are excluded.
  • Not all Salesforce object types, dashboards, or attachments are indexed.
  • Users may experience doc QA issues (e.g., difficulties with follow-up summaries or pasted links).

​
Update Salesforce Opportunity

​
Key Features

This action allows Glean agents to update specific fields on any Salesforce Opportunity that the user has access to. It integrates with your organization’s unique Salesforce schema, supporting both standard and custom fields.

  • Identify the Relevant Opportunity Record: Uses either AI suggestions or user confirmation from a dropdown list.
  • Preview Existing Live Values for Selected Fields: Displays current values and new suggestions.
  • Allow Users to Review and Edit: Enables users to review and modify any field value before changes are submitted.
  • Provide a Direct Link to the Updated Opportunity: Offers a link to the updated Opportunity on success or displays clear error messages on failure.

​
Usage Instructions

​
Set Up in Agent Builder

  1. In the Agent Builder, select the Salesforce update action by navigating to Select Step > Actions > By datasource > Salesforce > Update Salesforce Opportunity.
  2. Configure which Opportunity fields should be updated by selecting fields from your organization’s Salesforce schema using the dropdown.
  3. For each field, decide if the value will be filled by AI, chosen from a set, set as a fixed value, or provided by the user at runtime.
  4. Add instructions to guide how particular fields should be filled out.
  5. Save and deploy the agent.

​
Examples

  • Call Follow-up: After a customer call, an agent reads the call transcript and suggests updates to the Opportunity’s Stage, Next Steps, and Expected Close Date.
  • Batch Update: Reads changes from sheets or lists and batch updates information to Opportunities using agent looping.
  • Correction of Fields: If information changes or errors are detected (e.g., inaccurate Opportunity Owner or Amount), a user can quickly locate the Opportunity and update the relevant fields with guided assistance.

​
Best Practices and Troubleshooting

  • Identify the Opportunity in Previous Steps: Agents should have a search/read step to accurately identify the Opportunity. Skipping this may result in errors or poor matching.

​
Known Limitations

  • Not Supported in Background Agents: This action cannot be added to agents that run in the background.
  • Batch Processing: While batch processing is possible through loops, each update in the batch requires separate user confirmation—there is no “bulk confirm.”
  • No Automatic “Append/Overwrite” Detection: Users must specify text instructions in the step if new information should be appended to existing data in the field.
  • Support for Fields: Only fields the user has view/edit permissions for will be visible. Not all custom or complex field types may be supported in all releases.