For AI agents: a documentation index is available at /docs/llms.txt. Append .md to any page URL for markdown, or send Accept: text/markdown.
Import Salesforce Agentforce conversations into Agent Analytics
The Agentforce integration forwards Salesforce Agentforce sessions into Agent Analytics, including actions, tokens, and user feedback. A scheduled job in your infrastructure runs SQL over the Agentforce Session Tracing tables in Data Cloud, waits until each trace is complete, maps it to [Agent] events, and sends them to the Amplitude HTTP API. Amplitude then builds sessions, turns, and tool calls, and runs built-in signals.
This integration covers Agentforce agents with Session Tracing. Einstein Bots and agents outside Agentforce aren't covered.
Set up with your AI coding agent (recommended)
Paste this prompt into your AI coding agent, such as Cursor, Claude Code, Windsurf, GitHub Copilot, or Codex. The agent fetches the guide, builds the job, and starts with a dry run.
Connect our Agentforce sessions to Amplitude Agent Analytics.
Fetch https://raw.githubusercontent.com/amplitude/Amplitude-AI-Node/main/docs/integrations/agentforce.md
and follow "Part 2: Coding agent procedure". Start with a dry run.
The agent stops and asks for values it can't infer, such as which field holds your user ID. Review its changes, then verify with the steps in Get started.
What this means for you
- You run: A job on a schedule. Run it hourly.
- Your org needs: Enterprise, Performance, Unlimited, or Developer Edition with Agentforce and Data Cloud. Turn on Agentforce Session Tracing, and Audit and Feedback for tokens, models, and feedback.
- You need: An External Client App using OAuth client credentials with the
apiandcdp_query_apiscopes, and a run-as user with the Data Cloud User permission. - Your cost: Session Tracing storage and every query consume your org's Data Cloud credits. A large backfill is a real cost.
What carries over
| Data | Carries over |
|---|---|
| Turns and session grouping | Yes |
| Actions | Yes, as tool calls with name, inputs, outputs, and duration |
| Model, provider, and tokens | Yes |
| Thumbs feedback | Yes, as [Agent] Score events |
| Cost | No. Agentforce bills in Flex Credits, not per token. |
| Human agent messages after an escalation | No, they aren't in Session Tracing |
Don't assume the Einstein Trust Layer masks the stored traces the job reads.
Freshness and limits
- Delay: Session Tracing rows can take hours to days to land, so sessions arrive in Amplitude hours after they end, not minutes. This is a batch integration, not a real-time one.
- Incomplete traces: The job holds a trace until it's complete. After 72 hours, it forwards the trace and marks it incomplete.
- Rate limits: Salesforce's standard API limits apply.
Get started
Follow the Agentforce integration guide. It starts with a schema check against your org. Run a dry run on a window at least one day old, check Live Events and the session viewer, then re-run the window to confirm it adds no duplicates.
Guide last verified on October 6, 2026, against Salesforce documentation. Salesforce, Agentforce, and Data Cloud are trademarks of Salesforce. Amplitude isn't affiliated with or endorsed by Salesforce.
Was this helpful?