Which option to use

    What we recommend for a public website, a product behind a sign-in, and conversations containing personal details.

    4 min read · Updated 27 October 2026

    The right option depends on where the widget is, and what visitors ask about.

    A widget on a public website

    Visitors ask about your products, your prices and your opening hours.

    Leave all three settings off. Unverified addresses are enough for questions that do not depend on a visitor's history, and adding a code step would slow down every conversation.

    A widget inside a product people log into

    Everyone who reaches the widget has already signed in.

    We recommend Require identity verification with the identity token. Visitors are recognised without typing anything.

    Trust the typed email also works here, because your own sign-in controls who reaches the widget. Two things to weigh before choosing it: anyone who types an address will see the requests linked to it, and the session is kept on the device. Both matter more if people share machines, or if your team opens the widget on a customer's behalf.

    A widget where conversations contain personal details

    Previous requests may include account details, documents or other personal information.

    We recommend Require identity verification with the identity token.

    If you cannot add the token, use Ask for a code, so an address is confirmed before it is linked to a customer.

    What is changing

    At the moment, Ask for a code is asked before the chat starts, so every visitor sees it.

    This is changing. The code will be asked for when a visitor wants to see their previous requests, so a general question is not interrupted.

    If you are not sure

    Start with the stricter option. You can change it later, and each chat box is set separately.

    Was this helpful?