> ## Documentation Index
> Fetch the complete documentation index at: https://help.pipelime.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Agent settings reference

> Every field on the Settings tab, section by section, with what it changes and what happens if you leave it alone.

export const Screenshot = ({id, alt, caption, frame = 'browser', url, marks = [], lang = 'en', workspace = 'northwind-outbound'}) => {
  const src = `/images/screenshots/${lang}/${id}.png`;
  const chrome = url ? `app.pipelime.ai/${workspace}${url}` : 'app.pipelime.ai';
  return <figure className={`pl-shot pl-shot--${frame} not-prose`}>
      <div className="pl-shot__frame">
        {frame !== 'bare' && <div className="pl-shot__bar">
            <span className="pl-shot__dots">
              <span className="pl-shot__dot" />
              <span className="pl-shot__dot" />
              <span className="pl-shot__dot" />
            </span>
            <span className="pl-shot__url">{chrome}</span>
          </div>}
        <div className="pl-shot__media">
          <img src={src} alt={alt} loading="lazy" />
          {marks.map(mark => <span key={mark.n} className="pl-shot__mark" style={{
    left: `${mark.x}%`,
    top: `${mark.y}%`
  }} aria-hidden="true">
              {mark.n}
            </span>)}
        </div>
      </div>
      {caption && <figcaption className="pl-shot__caption">{caption}</figcaption>}
    </figure>;
};

The Settings tab is five collapsible sections (Identity, Channels, Message format, Goal and Advanced) over a save bar that appears at the bottom the moment you change something. Writing instructions live on **Agent Instructions**, and targeting and pace live on each lead source's own **Sources** tab; this page covers only who the agent is, who it writes from, how a message looks, what it drives toward, and who is assigned to it.

<div className="pl-availability">
  <div className="pl-availability__row">
    <div className="pl-availability__label">Where</div>
    <div className="pl-availability__value"><span className="pl-path">Outreach<span className="pl-path__sep">→</span>AI Agents</span></div>
  </div>

  <div className="pl-availability__row">
    <div className="pl-availability__label">Your role needs</div>
    <div className="pl-availability__value">Read access to AI Agents <code>read-campaigns</code>. Admin, Member and Viewer have it by default.</div>
  </div>

  <div className="pl-availability__row">
    <div className="pl-availability__label">To create or change</div>
    <div className="pl-availability__value"><code>create-campaigns</code> to add one, <code>update-campaigns</code> to change one, on top of the permission above.</div>
  </div>

  <div className="pl-availability__note">If you cannot find this in your sidebar, your workspace may have a custom menu configuration. Contact support and we will check it for you.</div>
</div>

Open any agent from the list, then its **Settings** tab. Every section starts collapsed (the left-hand nav is a map, not a menu) except two that open themselves the moment something is missing: **Channels**, when the agent has nothing to send from, and **Goal**, when the selected goal has no destination filled in. Both also raise a blocking banner at the top of the tab; the full catalogue of those banners is on [agent alerts and what to do about them](/en/ai-agents/agent-alerts-and-issues).

<Screenshot id="ai-agents/agent-settings-reference--sections" url="/ai-agents/128?tab=settings" alt="The Settings tab's left-hand section nav (Identity, Channels, Message format, Goal, Advanced) beside the Identity section expanded, showing the campaign name field and the Active/Paused switch" caption="Every section collapses to one header row until you (or a blocking problem) opens it: the nav on the left always lists all five." />

## Identity

| Field                                           | What it does                                                                                              |
| ----------------------------------------------- | --------------------------------------------------------------------------------------------------------- |
| **Campaign Name**                               | The name shown everywhere this agent appears: the agents list, its own header, alerts. Required.          |
| **Campaign is Active** / **Campaign is Paused** | A status line with an unlabelled switch beside it. The switch turns outreach on or off, and nothing else. |

The switch here is not a whole-agent kill switch. It writes one thing, whether outreach is on, and leaves lead sourcing running. Switch it off while a source is still going and the agent reads as **Outreach paused**, not **Paused**, and keeps spending credits finding companies nobody writes to. To stop both, use **Pause options** in the agent's header and choose **Pause everything**, or pause each source on the **Sources** tab. See [agent statuses and pause options](/en/ai-agents/agent-statuses-and-pausing).

## Channels

Three rosters, each its own card inside the section. **Email senders** and **LinkedIn** always show. **WhatsApp** shows only where WhatsApp is switched on for you, or once the agent already has a WhatsApp number attached, so losing the channel later still lets you detach what is there.

| Roster            | What it controls                                                                                                                                                                                                             |
| ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Email senders** | The **Choose specific senders** switch: off, the agent rotates across every email account in the workspace; on, it reveals a searchable, domain-grouped picker with **Select All**, **Clear** and **Only healthy** shortcuts |
| **LinkedIn**      | Which LinkedIn accounts the agent sends invitations and messages from                                                                                                                                                        |
| **WhatsApp**      | Which numbers the agent opens conversations from                                                                                                                                                                             |

Each roster's header carries a chip saying where it stands (**All senders**, a count of how many are selected, or **None**), so you can read all three at a glance. **Only healthy** is a replace, not a filter: it swaps whatever you had picked for every email account that is not in an error state.

Open **Channels** and the LinkedIn and WhatsApp cards each report how many of this agent's own leads carry a LinkedIn URL, or a phone number. Those two channels can only ever reach that subset. When none of the leads have a phone number at all, the WhatsApp card says so in amber instead of giving you a count.

Saving is blocked until one of the three has something in it: "At least one sender, LinkedIn or WhatsApp account is required". Leaving **Choose specific senders** off counts, because the agent then rotates across the whole workspace. But a workspace with no email accounts at all still trips the blocking banner, since rotating across an empty list sends nothing.

<Screenshot id="ai-agents/agent-settings-reference--channels" url="/ai-agents/241?tab=settings&section=channel" alt="The Channels section with Email senders open in specific-senders mode: the Choose specific senders switch on, a handful of mailboxes checked in the domain-grouped list, and a selected-count chip on the card header" caption="Specific accounts picked, not the default of rotating across everyone: the chip on the header tells you which mode you're in at a glance." />

## Message format

| Card           | What it does                                                             |
| -------------- | ------------------------------------------------------------------------ |
| **Plain Text** | Sends as plain text. Marked **Recommended**.                             |
| **HTML**       | Sends as HTML: images, branding and buttons, at a cost to deliverability |

A new agent starts on **Plain Text**. Below the two cards, a banner titled **Customize Campaign General Settings** points to Company Settings for follow-up delays and reply timing, with a **Configure Settings** button that deep-links there. Those apply to every agent in the workspace and are not set per agent. Full detail on [workspace-wide pacing](/en/ai-agents/campaign-pacing-settings).

## Goal

| Card             | Destination field | What you fill in                                                                                                                                                                                                                                                                                                                         |
| ---------------- | ----------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Meeting Link** | A meeting link    | A dropdown that opens on "Select a meeting link". A connected Calendly account lists its scheduling links; where **Meeting Links** is switched on for you, your own links are listed under **Your meeting links** with their duration. While nothing is selected, a second field beside it takes a custom meeting URL you type yourself. |
| **WhatsApp**     | A phone number    | A WhatsApp Business number, entered as a country plus a local number, plus an optional pre-filled message                                                                                                                                                                                                                                |
| **Website Link** | **Website URL**   | A URL the lead is invited to visit: accepts template variables, so each lead's link can carry its own company slug                                                                                                                                                                                                                       |
| **Custom**       | **Custom Goal**   | A free-text description of the outcome you want. No CTA link gets added                                                                                                                                                                                                                                                                  |

Switching the card clears whatever the other three goals had filled in: picking **WhatsApp** after typing a website URL loses that URL, not just hides it.

<Warning>
  A goal with no destination does not stop the agent. Outreach still goes out. It just never asks for anything, because the message is built without the link the goal exists to carry. That is why the tab treats it as an error rather than a warning: it opens the **Goal** section for you and shows a banner above the tab until you fill it in.
</Warning>

Where **Meeting Links** is switched on, a link with no calendar behind it is flagged twice: a **No calendar** chip in the dropdown, and, once it is selected, an alert saying leads cannot actually book a time until a calendar is connected, with a **Connect calendar** shortcut. See [share or embed a meeting link](/en/meetings/share-and-embed) for what putting a link here does to the emails the agent writes.

For **WhatsApp**, the country selector next to the phone field starts on Argentina, not your workspace's own country, so check it before typing the number. Once a number is in, **Test WhatsApp Link** opens that chat in a new tab with sample lead data substituted into the message, so you can see exactly what a lead lands on.

<Screenshot id="ai-agents/agent-settings-reference--goal" url="/ai-agents/241?tab=settings&section=goal" alt="The Goal section with Meeting Link selected and a meeting link chosen from the dropdown, showing its name and duration and the helper text explaining leads are sent straight to it" caption="Once a destination is chosen the field stops reading as missing, and the blocking banner above the tab clears on its own." />

## Advanced

| Field                                                                         | What it does                                                                                                 |
| ----------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------ |
| **Select Users**                                                              | Assigns every opportunity this agent generates to the teammates you pick, on a first-come first-served basis |
| **Audience tags**                                                             | The tags this agent's leads must carry. Changing them changes who the agent contacts                         |
| **Advanced editors** → **Open in Campaigns (Old)**                            | Opens the legacy campaign editor for this same agent                                                         |
| **Danger Zone** → **Delete outreach**                                         | Removes this agent's outreach campaign and stops its lead sourcing. Leads already found are kept             |
| **Danger Zone** → one **Delete** row per attached lead source, named after it | Stops sourcing new companies from it. Already-found leads are kept                                           |

Retagging can orphan a lead source. If a source's own tags no longer overlap the agent's new tags, saving stops on a **Tag change affects lead sources** dialog listing each one, and each starts on **Also add tag to lead source**: switch a source to **Detach** to let it go, then **Continue & Save**.

Both deletions ask you to type the name shown in the dialog to confirm. Deleting the outreach also offers a checkbox to delete the agent's lead sources with it (off by default, so they survive for reuse elsewhere), above a line counting how many of them are running right now and about to stop. **Delete outreach** needs the `delete-campaigns` permission and a lead-source **Delete** needs `delete-lead-generations`; without either, the whole **Danger Zone** is absent rather than greyed out, and without one of the two, only that button is.

## Saving and validation

A bar reading **You have unsaved changes** sticks to the bottom of the tab the moment anything on it changes, with **Discard** to revert everything and **Save** to submit. **Save** needs the `update-campaigns` permission; without it the button stays visible but inert, with a tooltip reading "You don't have permission to do this". It is the one control on this tab that greys out instead of disappearing.

Each row in the left-hand nav grows a small dot as you work: amber for a section holding unsaved edits, red for one with a problem: a field that failed validation, or the missing channel or missing goal destination that blocks the tab from the start. Saving with an error open forces every failing section open, whether or not you were looking at it, and a "Please fix the highlighted fields before saving" message appears.

<Note>
  Saving Settings never touches your Agent Instructions content. The tab carries `stage_instructions` and `linkedin_stage_instructions` through untouched on every save, even though there is no field here to edit them. That content lives entirely on the Agent Instructions tab.
</Note>

## Related

<CardGroup cols={2}>
  <Card title="Agent instructions" icon="message-square-text" href="/en/ai-agents/agent-instructions">
    Behaviour, offering, and per-channel replies and follow-ups: everything Settings deliberately leaves out.
  </Card>

  <Card title="Workspace-wide pacing" icon="gauge" href="/en/ai-agents/campaign-pacing-settings">
    Follow-up spacing, reply delays and one-contact-per-company: inherited by every agent, not set here.
  </Card>

  <Card title="Sender settings" icon="mail" href="/en/email-accounts/sender-settings">
    Limits, schedule and signature for one mailbox: what an account picked in Channels actually does.
  </Card>

  <Card title="Share or embed a meeting link" icon="calendar-days" href="/en/meetings/share-and-embed">
    Copy the public link, embed it or add a popup, and see what picking one as a Goal changes.
  </Card>
</CardGroup>