TwinsSalesforce
A Salesforce twin for testing AI agents.
Test agents that read and update the CRM. The same API and MCP calls, the same validation and permissions, and nothing changes in your real org.
A stateful copy of Salesforce, built for testing agents.
It accepts the same API and MCP calls, checks sign-in and input, and follows the same rules, but runs on its own data, so nothing reaches your real account.
- Accounts, contacts, leads and opportunities
- Field validation and required fields
- Stages and workflow rules
- SOQL queries
- Permissions, sharing and API limits
Every run starts from a company that already has history. We seed:
- Accounts with contacts and history
- Open opportunities at every stage
- Activities, notes and tasks
- Users with profiles
Use Salesforce in staging, without using production.
Validate every action
The twin checks what the agent sends the way Salesforce would: required fields, permissions, rate limits, malformed input. Mistakes fail here, not in your account.
Start from a realistic state
Describe the company and the situation in plain English. Drizz seeds the twin to match, and resets it to that state before every run.
Replay tasks across tools
The Salesforce twin shares its people, customers and history with the other twins, so a task that touches three tools runs the way it would at work.
Swap the address. Your agent code stays the same.
Replace the live Salesforce endpoint with the twin’s. Everything your agent already does keeps working, on data that is yours to break.
{
"mcpServers": {
"salesforce": {
"type": "http",
"url": "<twin MCP endpoint>",
"headers": { "Authorization": "Bearer <token>" }
}
}
}REST works on the same data. Both are reset with one call.
Connected to the rest of the company.
A deal in Salesforce has the email threads, meetings, Slack channel and Gong calls behind it, so an agent that updates the stage can be checked against what actually happened.
Read the playbook: Testing AI agents on Salesforce →
17 of 20 runs passed. 3 broke a rule.