Skip to main content
An agent is a great caller with nothing to do. A campaign is what gives it work. If the agent is the voice, the campaign is the plan behind the calls. A campaign answers the practical questions:
  • Who do we call?
  • When are we allowed to call them?
  • How often, and how far apart?
  • What number shows on their phone?
  • What happens after each call?
You set these once. Then the campaign feeds people to the agent, one at a time, and the calls just happen.

One agent, one campaign

Each agent comes with its own campaign, joined at the hip. The agent handles the talking. The campaign handles the logistics. You’ll set up the agent and its campaign together. A campaign belongs to one project, the sub-account whose contacts it calls, wherever that sub-account keeps them.

The pieces of a campaign

Each part has its own short guide:

Who gets called

How people join the call list, and how they come off it.

How it dials

The moment a lead arrives, on a schedule over days, or both.

When you may call

Calling hours, whose clock they run on, and after-hours leads.

How often you retry

Cadences, steps and the queue of leads waiting for their next try.

The phone number

The caller ID people see when you call.

After the call

What happens next based on how the call went.
The three timing settings, hours, mode and rhythm, are summarised together in When and how often you call. Campaign settings walks the settings screen control by control. Campaign types and capabilities explains what starts each kind of campaign and what the agent may do on it. The Power Dialer queue shows the people waiting for their next attempt. Testing and sharing covers the test form, web voice session and embed options. A master campaign extends one setup across linked projects with explicit rules for shared and local configuration.

It starts switched off

This is the most important thing to know: a brand-new campaign is paused. It will not call anyone until you turn it on. That’s on purpose, so you can set everything up, check it, and test it with zero risk. When you’re ready, you flip it live. See Go live.

In the API

The campaign comes back under campaign on Get agent and is changed with Update campaign. Fields you omit are left alone. The campaign’s name follows its agent’s name. A name sent on Update campaign does not last: it is set back to the agent’s shortly afterwards. To rename a campaign for good, rename its agent with Update agent, and the campaign follows straight away. A campaign keeps its own name in two cases: while it runs an active split test, and when another live campaign in the same project already has the agent’s name (compared ignoring case and spacing). config holds the settings that belong to one capability or one dispatch mode:
  • Booking: calendarId, collectAddress.
  • Transfer: transferNumber, transferToAssignedUser, advancedRouting with transferRoutes, and transferMessage.
  • Speed to lead: s2lDelaySeconds, s2lDelayHaltOnReply, s2lDoubleDial, callNowSkipTags, and the legacy s2lDelayMinutes.
  • Power dialer: cadenceId.
  • Callbacks: callbackDoubleDial.
  • Voicemail: voicemail and voicemailMessage, plus the optional voicemailDetectionPrompt. One message per cadence touch, on its last dial. Callbacks and inbound calls never leave one.
  • Tags the agent applies: appliedTag, confirmedTag, fallbackTag.
  • Testing: testFormFields, the stand-in lead fields a test call uses.
Two legacy aliases still appear: isActive for active and type for dialing. See Conventions.