Your AI agents can now support voice, SMS, and email from one coordinated profile, instead of being set up separately per channel. This article covers how that works and what you can configure per channel.
One Agent, Multiple Channels
Each AI agent (find yours under AI Agents > Agents) keeps a single identity, name, voice, and role, and you choose which channels it's active on. In an agent's configuration, scroll to Triggers, where you'll see up to three tabs:
Inbound Call: the phone number(s) routed to this agent. Inbound SMS: the number(s) it texts from. Inbound Email: the inbox(es) it monitors and replies from.
An agent doesn't need to be on all three. A leasing agent might handle calls, SMS, and email, while a receptionist agent might only handle calls. Whatever channels you connect, the agent maintains the same name, tone, and knowledge across all of them, so a prospect who calls and later texts is talking to the same "person." Talk to your account representative to have the expanded agents enabled!
Configuring Behavior Per Channel
While the agent's core identity is shared, each channel has its own settings, so you can tune behavior and safeguards differently for a phone call versus a text versus an email.
Call: connect one or more numbers, choose whether to hide the caller's number, whether to provide caller information to the agent, and set a fail-over number for when the agent can't handle the call. SMS: connect one or more numbers, set a prompt override (channel-specific instructions layered on top of the agent's main role), a max turns per conversation, and an idle timeout. Email: connect one or more inboxes, set a prompt override, max turns per conversation, and idle timeout, plus two controls worth calling out specifically below.
You can also apply an Email override or SMS override at the top of the Triggers section (toggle "Customize for Email" / "Customize for SMS") if you want channel-wide adjustments that apply regardless of which inbox or number the conversation came through.
Automatic Replies vs. Review-First Drafts
This is the one every team should look at before turning an email agent loose. In the Email agent settings, the Auto-Send Replies toggle controls what happens when the agent finishes composing a response:
On: the agent sends its reply automatically, no human touches it. Off: the agent still does the work, but the reply is left as a draft for your team to review and send.
Off is the safer starting point for a new agent or a sensitive inbox. Once you trust an agent's replies for a given inbox, you can switch it to auto-send.
There's also a Close Conversation When Agent Completes toggle below, which controls whether the conversation auto-archives the conversation once the agent considers its part done, versus staying open for a human to close manually.
Customer and Property Context
Under an agent's Knowledge section, Relevant Locations controls which properties and units the agent is allowed to talk about, you can scope it to specific properties, add a filter, or leave it open to everything published in your listing settings. This is what keeps an agent's answers relevant and appropriately scoped instead of pulling in information about properties it shouldn't be discussing.
The built-in Knowledge section on agents is voice-only. To connect specific knowledge articles for SMS and email, you'll link them as a capability.
