Skip to main content

​
Testing and Deploying

Before putting your agent in front of customers, use the built-in test page to verify it works correctly. Then deploy with a single code snippet.

​
Testing Your Agent

Every agent has a Test page in the dashboard. This gives you a live chat interface connected to your agent — identical to what your customers will see.

​
What to Check

  • Common questions — Ask the questions your customers ask most. Are the answers accurate and helpful?
  • Edge cases — Try vague or unusual questions. Does the agent handle them gracefully or give confusing responses?
  • Actions — If you’ve enabled actions (like ticket creation or appointment booking), test each one by using natural trigger phrases
  • Out-of-scope questions — Ask something unrelated to your business. The agent should politely redirect rather than making up an answer
  • Tone and style — Make sure the responses match the tone you configured
After testing, if answers aren’t good enough, go back and add more source material. More specific content leads to better responses.

​
Deploying Your Agent

Once you’re satisfied with testing, deploy your agent by embedding it on your website. You have two options. The chat widget adds a floating chat bubble to the bottom-right corner of your site. Customers click it to open a conversation. To embed the widget, copy the script tag from your agent’s Deploy section and paste it before the closing </body> tag of your website:
Replace YOUR_BOT_ID with your actual bot ID from the dashboard. The widget works on any website — WordPress, Shopify, Squarespace, Webflow, or any custom-built site.

​
Option 2: Iframe Embed

If you want to embed the chat directly into a page (instead of a floating bubble), use an iframe:
This is useful for dedicated support pages or help centers where you want the chat to be a visible part of the page layout.

​
Choosing Between Widget and Iframe

Both embed options connect to the same agent and use the same knowledge base and configuration. Choose based on where and how you want the chat to appear.

​
After Deploying

  • Monitor your agent’s performance from the Analytics dashboard
  • Update sources and Q&A pairs as your product evolves
  • Review conversations periodically to spot areas where your agent could improve