The Campaign Tester runs the full pipeline against your live config — real AI, real compliance, real context assembly. The only difference is no message is ever sent to a real number.
The Campaign Tester creates a test conversation linked to the campaign's current version. You type as if you were the contact — the AI responds exactly as it would in production. The right panel shows the full turn detail: directive, reasoning, confidence, and the context the AI received.
The Campaign Tester is the right place to validate question flow, branch logic, tone, and edge cases — before any real contact receives a message.
Walk through the full question sequence to confirm questions appear in the right order, branch instructions fire correctly, and the AI handles both positive and negative answers as expected.
How it works: Branch shortcuts let you jump to specific question paths without manually simulating the full conversation.
Test how the AI handles off-topic messages, repeated objections, out-of-hours booking requests, and explicit disengagement — before any of those scenarios reach a real contact.
How it works: The AI's reasoning field shows exactly why it chose each directive — making unexpected behaviour easy to diagnose.
Read the AI's actual outbound messages to confirm tone matches the campaign's intended voice, filler phrases are absent, and character limits are respected throughout the conversation.
How it works: Fallback turns — where the AI's response was substituted — are flagged visually in the turn detail panel.
The test conversation is always created against the campaign's current version. When you reset, the new conversation picks up any config changes made since the last test — no cache to clear.
How it works: current_version_id is captured at conversation creation — the tester always reflects the live config.
We'll set up a Campaign Tester session and walk through the full conversation flow with you.