Test scenario
The thread in this column didn’t create this file. Your spec is on the right—run a Live test, or start a new chat so the agent can help you improve the scenario, then Regenerate for fresh JSON.
Nothing to show yet
When your test design is ready, press Generate next to the send button, or open a saved file from Scenarios in the header. A readable summary of the plan will appear here.
Nothing to show yet
After you generate a scenario (or open one from the library), this tab shows a short sample conversation when available so you can check that persona, steps, and test data fit together. It is only an illustration, not a guarantee of how the bot you’re testing will reply.
Illustrative sample dialogue
This is a simulation only. Replies from the bot you’re testing may differ. We cannot know exactly what production will return.
Use this preview to confirm your test scenario is structured correctly: persona, steps, and test data line up so the run is ready to execute.
Nothing to show yet
The downloadable test scenario JSON is available after you generate or open a file from the library. You can copy, download, or save to disk from here.
Load a scenario first
Live test runs your scenario against the bot you’re testing (via Botium) and the objective-testing API. Press Generate in chat, or open a file from Scenarios, then use Run test next to the send button.
These messages are about the connector / API run only. They are not part of your scenario chat on the left.
Step 1: click a connector card to select it. Step 2: press Run test. Empty areas under “Live conversation” / “Objectives” fill in once the run starts.
Tip: Test connection on a card checks the bot before a full run.
What is Debug State?
This is the builder’s working state: structured fields gathered from your conversation (chatbot context, personas, objectives, test data, and flags such as IVR or blocking steps).
It is not the same as the Generated Scenario JSON on the JSON tab. That file is produced when you press Generate or Regenerate. Use this view to see what the assistant has captured so far, or to troubleshoot if something looks wrong.
It updates after each assistant reply and when you open this tab. You can skip this during normal scenario design unless you want extra detail.
State not yet available.