Conversations and handoff

Conversations in the dashboard holds every shopper conversation with your AI storefront, your widget and your ChatGPT and Claude app - and the requests that wait for your reply. Human handoff is how a shopper reaches a real person when the AI should step aside: they leave their email in a short contact form, you get notified, and you reply by email straight from Conversations. The AI sells; it does not trap people - and for shoppers, knowing a human is reachable is what makes an assistant trustworthy.

Reading conversations

Conversations are listed under All, Waiting and Archived. Open one to read the whole exchange - what the shopper asked, which products the assistant showed - and mark it answered when you are done. You can reply once the shopper has left an email address.

When the contact form appears

In your AI storefront and widget, the form appears in four situations - three automatic, one manual:

• The shopper asks for a person. "Can I talk to someone?" in any supported language opens the form instead of a deflection.
• The assistant cannot answer. When a question is beyond the catalog and store knowledge, the storefront offers the human route rather than guessing.
• A purchase is blocked. If a shopper wants to buy but checkout cannot complete, losing them silently is the worst outcome - the form catches exactly that moment.
• The shopper opens it themselves. The contact option is always reachable in the storefront, without needing a trigger.

What the shopper sees

• A short form: email (required), an optional name, and their message - prefilled with context from the conversation where it helps.
• The form renders in the shopper's language (24 languages), including the consent line.
• A consent line stating that the store may contact them about their enquiry. The exact wording and time of consent are stored with the request - not just a checkbox flag - so consent stays verifiable later.

What you receive, and how you reply

• A notification email with the shopper's message, email, and the reason the handoff was triggered. It goes to the address you set under "Send contact requests to" in the storefront studio - or to your account email if you have not set one.
• The conversation lands in Conversations, marked as waiting, with the shopper's name and country. The full chat history sits right there, so you see what the shopper already asked before writing a reply.
• You reply from Conversations. The reply editor supports bold, italics, lists and links; your answer reaches the shopper as an email branded with your store's name, and the shopper can answer it directly - their reply goes to your address, not through AgenticCart.
• Handled requests leave the waiting list automatically once you have replied; conversations can also be archived and restored.

Turning it on

• AI storefront and widget: open the storefront's Studio and switch on Human handoff under Actions. Optionally set "Send contact requests to" if notifications should go somewhere other than your account email - a shared support address, for example.
• ChatGPT and Claude: on the ChatGPT & Claude page, switch on Messages from shoppers. With the shopper's consent, the assistant can then send you their question; it arrives by email, your reply goes to the shopper, and it shows up in Conversations. Reviewers at OpenAI and Anthropic may send a test message.
• That is all. The triggers, languages and consent handling work without further configuration.

Good practice

• Reply from Conversations, not from memory: the conversation context is attached there, and the reply is recorded with the conversation.
• Only contact shoppers about their enquiry. The consent covers the enquiry - it is not a newsletter opt-in, and the notification email says so.
• Watch the waiting counter. A handoff is a shopper at their most convertible moment - they asked for you by name. Speed matters more here than anywhere else in the funnel.

Next steps

See customize your AI storefront for the other Studio controls, ChatGPT & Claude app for your store's app, and insights for what shoppers ask across all conversations.