| name | pendo-agent-analytics-agentforce |
| description | Deploy and verify Pendo Agent Analytics for Salesforce Agentforce — capture every Agentforce conversation turn (prompt and response) into Pendo. Use when the user wants to deploy this repository, set up Agentforce-to-Pendo conversation tracking, or configure the Data Cloud (or Custom Action) integration. |
| compatibility | Requires the Salesforce CLI (`sf`) authenticated to the target org, shell network access, and — for the recommended Data Cloud approach — a Data Cloud-provisioned org. |
| license | Apache-2.0 |
Deploy Pendo Agent Analytics for Agentforce
You are running a guided, interactive deployment of this repository for the user.
- Read
AGENTS.md at the repo root — it is the orchestration guide (preflight, what to automate, where to pause for the human, verification, and guardrails).
- Read the Data Cloud Approach — Deployment section of
README.md for the exact, authoritative steps.
- Execute the deployment following those two files:
- Do the automatable steps yourself (deploy the minimal set, run tests, set
Pendo_Config__c, read ssot__AiAgentApiName__c, create the Pendo_Target_Agent__mdt record, deploy the Flow, run the smoke test).
- Pause to ask the user for the Pendo API Key + Agent ID + region, and which agent(s) to track.
- Guide the user through the human-only UI steps: enabling Session Tracing (then the 15–30 min wait) and activating the Flow.
- Verify with the anonymous-Apex smoke test and Pendo Raw Events.
Default to the Data Cloud approach (recommended). Never echo or commit secrets — they go only into Pendo_Config__c. Do not deploy the example agents in bots//genAiPlannerBundles/. Confirm before any outward or irreversible action.