When your organization is set up on a paid plan, Laureo installs its preset agents on the Agents page and turns them on for you. Each one arrives on Ask before acting and starts on its own: when its work shows up (a form lead, a new company, an opened quote or an email from a contact) or, for quiet deals, every weekday morning. At that level: Looks things up and prepares drafts. Asks before changing, sending or deleting anything. You can watch what they propose on the Agents home from day one.

An agent with nothing to act on does not run and uses none of your AI credits. Until its first run, its card reads Not run yet with a line under it saying what it is waiting for.

You are in control of every one of them. From the Agents section of the Agents home, an admin or the marketing role can pause or resume any of them. To change how one behaves, open it and edit it in place: the Build tab shows the published version read only until you choose Edit as draft, and nothing you change runs until you publish it.

Starts when someone submits one of your published forms. It reads the submission and the company it came in with, finds who owns the person who submitted the form, and drafts a short first reply to the person for you to review and send. It proposes an owner for the lead where one is needed, and a follow-up task for that owner. It never sends the reply itself.

Runs a risk scan of your open deals each weekday morning, flags the ones that have gone quiet or slipped past their expected close date, and proposes one concrete next step on each flagged deal. It then drafts a short follow-up to the deal contact that refers to the last real interaction. It prioritises high-value and late-stage deals, includes dormant accounts that once showed interest, and skips deals whose quiet period is explained by a future close date.

Starts when a company is added. It looks for an existing company that is the same one, and when it finds a likely duplicate it proposes merging the two, so the history ends up on one record. It checks each company as it arrives rather than sampling your records once a week, and it never merges on its own.

Starts when a client opens a quote you sent and has not accepted it. It reads the quote, the company and people on it and their recent activity, then drafts a short follow-up to the client for you to review and send, and, when the quote expires within seven days, proposes a follow-up task for its owner. It never sends the email and never changes the quote.

Starts when an email from a person already in your CRM arrives in the mailbox of the teammate it runs as, and reads that mailbox only. It works out what the email calls for and proposes it: a link to the open deal it belongs to, a task for a commitment other than a reply, a ticket for a support problem when that person has none open, or a new contact for someone new in copy or in a signature. It picks a deal only when exactly one open deal matches or the email names it; otherwise it records the email on the person and says so in the run. It drafts nothing and sends nothing.

Ticket Triage is one of the defaults, but it is not an agent you edit on the Agents page: it is how the Service AI meets each new ticket from a customer, so you switch it on and off in Service settings. It sorts each new customer ticket, finds the matching help article and leaves a first reply draft on the ticket for your team to review. It sets the priority of a new ticket itself, without an approval card, and it never sends a reply, resolves a ticket or assigns one to a person.

See Triaging Tickets in the Inbox for working the queue it sorts.

If your workspace had Daily Brief or Collections Chaser, its card on the Agents home now points to where that work happens, with one link and no switch.

A preset agent is a starting point, not a fixed product. Open one, choose Edit as draft, and change its brief, its schedule, what it is allowed to do, or how much it may do on its own. The agent you edit is the one that was installed for you, so there is no copy to keep in step; the live agent keeps running the published version until you publish your changes, and every publish is a version you can restore. See Building Your Own Agent.