Bot Testing and Preview Mode
Bot Testing & Preview Mode
Before publishing a new bot or workflow change, use Velaro's built-in testing tools to verify that your bot responds correctly, routes properly, and handles edge cases gracefully. Testing catches issues before real customers encounter them.
Bot Preview Panel
Every bot configuration page has a Preview button (bottom right). Clicking it opens a live chat window connected to your bot in a sandboxed test environment. Changes you make to the bot configuration are reflected in the preview immediately — you do not need to save and publish first.
The preview panel shows both the customer-facing conversation and a Debug Sidebar that displays:
- Which workflow step is currently active.
- Variable values as they are collected.
- Which skill was invoked and what data was returned.
- Confidence scores for AI responses.
- Any errors thrown by skill calls.
Test Scenarios to Cover
Always test these scenarios before publishing a new bot:
| Scenario | What to Check |
|---|---|
| Happy path | Visitor follows the intended flow from start to resolution. |
| Escalation request | "Talk to a person" — bot hands off cleanly with context preserved. |
| Off-topic question | Bot handles questions outside its scope gracefully (clarifies scope, redirects, or escalates). |
| Ambiguous input | Short or unclear messages ("yes", "no", "help") — bot asks a clarifying question rather than guessing. |
| Empty knowledge base hit | Question the knowledge base cannot answer — bot acknowledges it doesn't know rather than making something up. |
| Skill failure | Simulate a CRM or integration error — bot should gracefully fall back rather than crash or loop. |
| After-hours | If your bot has business hours logic, test it both during and outside hours. |
| Multilingual | If translation is enabled, test with a message in the target language. |
Testing Specific Channels
The preview panel simulates web chat by default. To test channel-specific behavior:
- SMS / WhatsApp: Use a real test number configured in Settings → Channels. Send messages from your personal phone to the test number.
- IVR: Use the Test Call button in Settings → IVR to call your own phone and walk through the IVR flow.
- Email: Send a test email to your connected support address and observe how the bot responds in the inbox.
Staging vs. Production Bots
Velaro does not have a separate staging environment, but you can manage this with workflow versioning:
- Create a test deployment on a staging page or internal URL with a separate Velaro deployment configured.
- Assign a draft workflow to the test deployment.
- When satisfied, promote the workflow to production by assigning it to the live deployment.
Regression Testing After Changes
When making changes to an existing live bot, use the Conversation History to review the last 50 conversations as test cases. For each conversation, verify that the updated bot would have given the same or better response. Pay attention to conversations that were escalated to a human — these are often edge cases your bot mishandled previously.
Reviewing Test Results
After a test period, go to Reports → Bot & AI → Test Sessions to see a summary of all preview sessions. This report highlights which steps had the most drop-offs, which skills failed most often, and which inputs triggered fallback responses — all pointing to where further tuning is needed.
Was this article helpful?