The thread, not just the message
Incoming messages and the conversation history around them. So a thread can be shown complete, instead of as one message with no context.

Connect the mailbox you already use. Dolphin AI drafts the reply, and nothing goes out automatically unless you switch it on.

Two direct connections and one that works with everything else.
Connect a Gmail or Google Workspace mailbox directly. Sign in once and threads start arriving with their history intact.
Connect an Outlook or Microsoft 365 mailbox the same way. Same inbox, same history, same handling.
Forward your mailbox to AskDolphin and it works exactly like the other two. Your provider doesn't need to be on a list.
It isn't a switch you flip on.
It's something you set up, assign and pause.

Name it, pick its mailboxes, and choose the address it replies from. Run one, or several that work differently.

Switch its Data Sources on or off: your pages, products and Q&As, or nothing. Add reply rules in plain English.

Add your own footer note. AskDolphin adds a line identifying every AI reply, so nobody is left guessing.
And when it is on, you still choose which emails it answers.

Every agent starts with automatic replies off. Dolphin AI drafts; your team edits and sends, or doesn't. Many never switch it on.
New customer emails, only threads a person has already joined, or both. Pick the second and it never starts a conversation on its own.
Dolphin AI won't reply to obvious spam, promotional email, bounces or other automated messages. No arguing with another robot.
Every agent has a status, and one switch turns them all off across every mailbox. Nothing you set up gets deleted.
Verify your domain and every reply leaves with your address on it.

Add your domain and verify it, then send from any address on it. Most teams skip this step.
Create addresses on that domain and pick which one each agent replies from. That is what customers see.
Mail from a verified domain is more likely to arrive where it should. Ten minutes that pays off.
Not buried three pages into a policy.
You grant access explicitly, through your provider's own sign-in or by setting up forwarding yourself. Nothing is accessed before you do that.
Incoming messages and the conversation history around them. So a thread can be shown complete, instead of as one message with no context.

Those messages, inside your AskDolphin inbox. Visible to the people on your team with access to it.

Replies you or Dolphin AI compose. Going out from the address you chose.

Something else?Ask us.
Gmail and Google Workspace connect directly. Outlook and Microsoft 365 connect directly. Anything else connects by forwarding your mailbox to AskDolphin, which means your provider doesn't have to be on a list for this to work.
Not unless you turn that on. Automatic replies are off by default on every agent. Until you switch them on, Dolphin AI writes drafts and your team decides what sends.
Yes. Choose new incoming customer emails, only threads a person has already joined, or both. Set it to the second and it never starts a conversation on its own.
No. Dolphin AI won't reply to obvious spam, promotional email, bounces or other automated messages.
Yes. Each agent has its own name, its own mailboxes, its own reply instructions and its own purpose, so a sales address and a support address can behave completely differently from the same workspace.
No. You keep the address you already publish, and replies go out from it once your domain is verified.
Yes. AskDolphin adds a short line identifying an AI reply automatically, and you can add your own footer note above it.
It reads incoming messages and their history so threads can be shown in full, displays them in your inbox, and sends the replies you or Dolphin AI compose. That's the whole of it, and access only starts after you grant it.
Nothing sends until you say so, so there's no risk in looking.
14 days, every feature.