How to Customize Your Chatbot Widget’s Branding and Welcome Message

Before you start

Open Customization using a workspace admin account. Have your brand colors and a clear logo image ready. Decide which visitor questions the assistant can reliably answer before writing its greeting or starter questions.

The page includes a Live Preview for checking the appearance as you edit. Use it for an initial review, then check the installed widget on your website: the preview does not replace a real conversation or mobile layout check.

1. Set colors and upload your logo

1. Under Colors, set Primary Color (header background) using the color picker or a hex value such as #16324F.
2. Set Accent Color (user bubbles, send button).
3. Set Text Color (header text) and check that it is easy to read against the header background.
4. Under Logo, select Upload Logo and choose an image within the displayed 2 MB limit.
5. Wait for the upload to finish and review the image in the preview.

Choose colors for legibility as well as brand consistency. A detailed logo may become difficult to recognize in a small header; use a simpler approved version when available.

Reset Colors restores the color controls to their defaults. Review the result before saving.

2. Name the assistant and explain what it can help with

Under Assistant, edit Assistant Name and Welcome Message. Use a name that makes the assistant’s role clear, such as “Acme AI Assistant.”

A useful welcome message names a few supported tasks. For example: “Hi, I’m Acme’s AI assistant. Ask me about our services, coverage areas, or how to request a quote.” Use this example only if your sources support those topics.

Avoid promising instant human support, confirmed bookings, or account changes unless your workflow supports them. Editing a welcome message changes the introduction; it does not add business facts to the knowledge base or define every generated reply.

For answer accuracy, use the answer improvement guide alongside these appearance settings.

3. Add useful starter questions

Under Starter Questions, use Add question to enter short questions a visitor might ask. Edit existing entries or choose Remove for questions that no longer fit.

You can also select Suggest from knowledge base. Review the suggestions before using them. Each starter should point to an answer your current sources can support.

• “Which service is right for a small office?”
• “Where do you provide your service?”
• “How do I request a quote?”

These are examples for a service business. Replace them with your own visitor tasks, then ask each question in Test Your Agent to check the answer.

4. Choose the position and launcher label

Under Position, choose Bottom Right or Bottom Left. Pick the corner that fits your website’s other controls, then verify it on desktop and mobile.

In Launcher Button, enter a short label such as “Ask AI.” Leaving this field blank gives an icon-only button. A clear label can explain the purpose before someone opens the panel.

Under Proactive Greeting, you can edit the Bubble message and delay. The dashboard describes a greeting that appears after the delay or after scrolling a third of the page, whichever comes first. Set the delay to 0 to turn it off. A blank bubble message falls back to the welcome message.

Use a short invitation tied to a supported task. Check that the bubble does not cover important page content.

5. Save and verify the live widget

1. Select Save Changes and confirm the saved state appears.
2. Refresh a page with your installed widget and open the chat.
3. Check the logo, colors, name, greeting, starter questions, and launcher.
4. Ask a starter question and inspect the answer.
5. Repeat on a phone. Check that the button and panel do not obscure navigation or enquiry controls.

If changes are missing, confirm they were saved, refresh the page, and check that the site uses the embed snippet from the workspace you edited. If the widget itself is absent, follow the installation troubleshooting steps.

If you need help, contact support with the affected page, setting, and expected result. For collecting enquiries after a conversation, continue with the lead capture guide.