You are here:
Timesheets and Labor Cost Optimization Requirements
Before you set up timesheets and labor cost optimization with AI tools, confirm that your org has Field Service installed and the required licenses, features, and edition. You also need admin credentials, the Salesforce CLI, an AI coding agent, such as Claude Code or Agentforce Vibes, and at least one technician user and one admin user.
Before you begin, see AIforce documentation.
Before
Prerequisites
- Set up your org for the Salesforce Hosted MCP Servers.
- Activate the Headless 360 MCP Server.
- Connect your coding agents.
- Access to a Salesforce Energy & Utilities org with all necessary licenses and Field Service installed.
- Login URL and admin credentials for the target org.
- Access to an AI agent tool configured to run the setup skill (for example, Claude
Code or Agentforce Vibes running the setup skill that executes the Setup Operation
Recipe (SOR) files).
Note For Claude Code, install the Salesforce Development plugin, which provides Salesforce-specific skills and org context through hosted Model Context Protocol (MCP) servers. - Salesforce CLI
- Permissions to authenticate the CLI tool against the target org.
- At least one technician user (Standard User profile)
- At least one Admin/Approver user (System Administrator profile)
Products and features
- Field Service
- Timesheets and Labor Cost Optimization
- Field Service mobile app
Editions
- Developer
- Enterprise
- Performance
- Unlimited
Suggested Technology and Experience
An AI agent such as Claude Code or Agentforce Vibes runs this setup from a command-line interface. The agent connects to the org through a Model Context Protocol (MCP) server and uses the Salesforce CLI to authenticate to the org. The source doesn’t call out a required model or version. When a headless metadata call isn't available, the workflow falls back to the Salesforce CLI. See Agentic Tips for the fallback prompt.
The workflow relies on these components:
| Component | Purpose | Required? |
|---|---|---|
| Coding agent (Claude Code or Agentforce Vibes) | Runs the multi-phase setup from a command line | Yes |
| Headless 360 MCP server | Exposes the headless setup/metadata routes the agent calls | Yes |
Headless metadata skill (SOR steps that call the
/headless/metadata route) |
Applies setup metadata during the workflow | Yes |
Timesheet recipe
(deploy-timesheets-end-to-end) |
Drives the timesheets and labor cost optimization setup path | Yes |
| Salesforce CLI | Authenticates the agent to the org (sfdx org login web) | Yes |
This solution is for a Salesforce admin (System Administrator) who is comfortable running commands in a terminal and validating configuration in Setup.
Agentic Tips
- Recommended prompt: Start the setup with a setup time sheet end-to-end .
- Approvals: When the agent locates the recipe, it asks you to confirm, type Yes to proceed. During setup, the agent prompts you to choose how permissions are assigned (Permission Sets or Permission Set Groups), to provide the technician, supervisor, and admin user details, to confirm optional features (such as Mobile Crew Lead Time Management), and to set field values (such as the End Time Rounding Threshold in minutes).
- Authentication: The agent sometimes prompts you to authenticate the MCP
server against your org. In a separate terminal window, run
web to log in through your browser, then return to the agent terminal.sfdx org login - Context to provide: The user details for the Time Logger (technician), Time Approver (supervisor), and admin; your permission-assignment preference; any optional features to turn on; and any field values to change. Never paste credentials into a prompt.
- Interaction pattern: The agent runs in three phases — assign permissions, turn on Labor Cost Optimization settings, then configure the Field Service mobile app — pausing for your input at each decision point. Afterward, verify the result in the Salesforce UI with the steps in Validate the Timesheets and Labor Cost Optimization Setup in Salesforce.
- Model: The source doesn't require a specific model. State a model only if testing shows that the workflow requires or performs best with one.
