Campaigns live under Manage → Campaigns. That page isn’t in the default sidebar for the Agent Builder role. Add it with Customize sidebar, or press Cmd/Ctrl+K and search for it. See Navigating Studio.
Launch Checklist
The Launch button stays disabled until every gate is met. Hover it to see the checklist:- Campaign channel selected — set Channel to Voice on the Setup tab. A new campaign starts with no channel chosen, so this is the first item on its checklist. It drops off the list once a channel is set.
- AI Agent assigned
- Schedule configured — timezone plus at least one active day
- Contacts uploaded — at least one contact
- Number assigned to agent — the campaign’s AI agent has at least one phone number that can place campaign calls: a number bought in the Studio, or a BYO number on a Verified SIP trunk. When the agent’s only numbers are BYO numbers that aren’t on a Verified SIP trunk, this item reads Agent’s BYO numbers on a verified SIP trunk instead. See Campaign Setup and Phone Numbers
- SAN configured — required when DNC scrubbing is enabled (the default); you can save it in Campaign account settings or give it to the Squawk Co-pilot to save for you. See Compliance & DNC
Compliance Attestation
Launching from a campaign’s own page shows a Confirm Compliance dialog. It appears whether you use the header Launch or Relaunch button or the Launch Campaign button on an empty Live Monitoring tab. You must confirm you have the legal right to contact the recipients and that the campaign complies with applicable regulations, including the TCPA. When DNC scrubbing is enabled, the attestation also covers the SAN saved on your account. Confirming records the attestation in the campaign’s audit trail. The Launch and Relaunch icons on each row of the Manage → Campaigns list start the run directly and don’t show this dialog. Your consent and TCPA obligations are the same however you launch. To go through the confirmation step, launch from the campaign’s own page. After launch you’ll see “Campaign launched successfully (run #N)” — or “Campaign queued — dialing starts …” if you launched outside the schedule window.Runs
- Each launch creates Run #N with immutable snapshots of the schedule, dialing settings, agent, and the agent’s phone numbers taken at launch time — so you can always see which settings, agent and numbers a past run used, even after editing the campaign. The agent snapshot records which agent the run used, not what that agent says: each call uses the agent’s published version if it has one, and otherwise its latest saved edits, read when the call connects, so a Publish during a run changes the calls that follow. See Publishing Your Agent. The phone-number snapshot is a record only: the dialer reads the agent’s numbers on every dial, so adding or removing a number mid-run takes effect from the next dial.
- Alongside those numbered runs, a campaign has at most one permanent on-demand run, created by the first on-demand API call. It is deliberately excluded from the Runs tab, from Live Monitoring and from campaign list statistics, so your CSV run reporting stays comparable. On-demand activity is reviewed from the campaign’s On-Demand tab, and its calls appear in Call History like any other.
- Only one run can be active per campaign. Launching while a run is active returns “An active campaign run already exists.”
- Pause suspends dialing (in-flight calls finish); Resume continues (after a low-balance pause, add funds first — see Troubleshooting); Stop ends the run — remaining contacts are skipped and the run stays on record for review.
- Stopped, completed and failed voice campaigns can be Relaunched, creating the next run against the current contact list and settings.
- Run numbers are shared with the on-demand run. If a campaign’s first on-demand call came before its first CSV launch, its first CSV run is Run #2, and the Runs tab can show gaps in the numbering.
Live Monitoring Tab
The Live Monitoring tab is a real-time dashboard for the latest run. A pulsing green Live pill shows while the stream is connected; updates arrive within seconds of each dial event, with a Refresh button as backup.Run overview
Run number and status, channel, a live banner with the number of active calls right now, and chips for the agent (from the run’s snapshot), schedule timezone, and start/end timestamps — rendered in the run’s schedule timezone.Progress and KPIs
- Progress — percent complete with processed / pending / total counts.
- Pending — contacts not yet dialed or queued for their next attempt.
- In progress — dials or live calls in flight.
- Connected / completed — successful connects.
- Failed — errors, invalid numbers, or carrier failures.
Campaign Dial Status buckets
A collapsible breakdown of every outcome and queue state. Voicemail, Wrong number and Callback requested are conversation outcomes the AI agent reports during the call; the rest are queue states set by the dialer.How a campaign call ends
Every conversation outcome is recorded through the same end-of-conversation path, so it lands on the contact exactly once whether or not the agent says anything first. What the contact hears depends on the outcome:Voicemail, Wrong number, DNC request and Callback requested are reserved outbound-campaign outcome names. They drive the campaign follow-up — rescheduling a callback, marking the contact, or adding it to your do-not-call list — so an agent’s own custom outcomes cannot reuse those names.
Contacts table
A filterable, per-contact view of the latest run. The columns are Phone, From (the number the contact’s most recent call went out from), Status, Attempts, Reason, Last attempt, Retry after, Avg duration, Zipcode and Created. Click a heading to sort by Phone, Status, Attempts, Last attempt, Zipcode or Created. The other columns don’t sort. Click any row to open the contact’s dial history. Consent details aren’t columns in this table. You’ll find them in the contact’s dial history, under Consent status. The Dial status filter offers All statuses, Pending, In progress, Retry (incl. frequency cap), Completed, Callback requested, No answer, Busy, Failed, Blocked, Voicemail and Wrong number. The Search this page box matches phone, zipcode, status, reason and custom-field values, but only among the rows loaded on the current page — 20 by default, or 10 or 50 if you change the page size. To look further, page through the table or narrow it with the Dial status filter. Export downloads the contacts as a CSV, up to 50,000 rows. It follows the Dial status filter you’ve set, not the search box. If more than 50,000 contacts match, the file ends with a row reading “Export truncated at 50,000 rows”. The CSV columns come out in the same order as the table, under technical headings:phone, last_from_number, dialing_status, attempt_count, reason, last_attempt_at, retry_after, scheduled_at, avg_call_duration_seconds, postcode and created_at. Any custom fields from your contact list follow, appended alphabetically. scheduled_at isn’t a column on screen. It holds the requested call time for an on-demand call, so it’s empty for contacts from a CSV upload. The reason column holds the same text as the table’s Reason column, left empty where the table shows a dash. Downloads are named after the campaign and run, for example summer-outreach-run-3-contacts.csv.
The Reason column explains why a contact failed, is waiting, or was blocked, in plain English. For a contact blocked before it was dialed, it’s the compliance check’s own sentence — for example “Contact is missing prior express written consent and must not be dialled” or “Contact has opted out of voice calls and was added to the internal DNC list”. Long reasons are cut off in the table; open the contact’s dial history to read them in full. The run detail page uses the same wording, so the two views agree.
Per-contact drill-down
Click any contact to open its dial history drawer:- A summary at the top: Attempts (this contact), the contact’s current status, Retry after when another attempt is scheduled, and a Blocked or Queued line with the reason when the contact was held back.
- Consent status: Has PEWC and Voice consent, each shown as Yes or No, plus the PEWC date when the contact has PEWC and a date on file.
- Every dial attempt, newest first, with its call status, Start, End, Duration and the From number the call used.
- The TCPA decision for each attempt — Call placed, Queued for later, or Blocked from dialing — with the reason under Detail, plus a Failure reason when the call failed.
- The Attestation the call received (STIR/SHAKEN), such as A, B or C, and the Call SID with a copy button.
- For answered calls: expand View transcript on the attempt to see the conversation Summary and the Transcript. The transcript opens with the conversation only. Switch Action Logs on to show the actions the agent ran alongside it (off by default). When the call was recorded, a Call Recording bar has a Download button.
Runs Tab
The Runs tab lists every run as a card: run number, status, launch-to-end timeline with duration, and problem-status counts — colored dots with counts for blocked, failed, wrong number, voicemail, no answer, and busy — so you can spot troubled runs without opening them. Clicking a run opens the run detail page: summary tiles (total contacts, dial attempts, average call duration), a by-call-status breakdown, a searchable contacts table with the same drill-down drawer, and the run’s frozen Schedule, Dialing settings, and Caller ID snapshots. The Caller ID (snapshot) card shows Source as “Agent’s assigned numbers”, how many Numbers available the agent had at launch that could place the run’s calls, and a View numbers (N) list. Runs launched before 14 October 2026 keep the list they started with, which can include BYO numbers that aren’t on a Verified SIP trunk. Runs launched when campaigns still called from outbound number pools keep their original pool details, with the pool’s name next to Source. The run detail contacts table is similar to Live Monitoring’s, with a few differences:- The columns are Phone, From, Status, Reason, Attempts, Total duration, Last attempt, Zipcode and Created, plus a Details button on each row that opens the drill-down drawer. There is no Retry after column, and Total duration takes the place of Avg duration.
- You can sort by Phone, Status, Attempts, Last attempt, Zipcode or Created.
- The search box looks across every contact in the run, not just the current page.
- The status filter offers the same options as Live Monitoring’s Dial status filter. Export works the same way, and the exported CSV has the same columns.
Compare Campaigns
From the campaigns list, click Compare and select 2–5 campaigns to see their latest runs side by side: total contacts, dial attempts, connect rate, conversion rate, average and longest durations, and outcome counts — with the best value in each row highlighted. That is the highest value, except for No Answer, Busy, Failed, Blocked and Pending, where the lowest is highlighted. How the figures are worked out:- Dial Attempts counts every dial attempt recorded for the run, including attempts that were queued or blocked before a call was placed.
- Connect Rate is completed calls divided by all dial attempts.
- Conversion Rate is contacts that reached Completed divided by total contacts.
- Avg Duration and Longest Call include completed calls only.
- A campaign’s latest run is the one launched most recently. A campaign’s on-demand run counts as launched when the first on-demand API call creates it, so if that came after your latest CSV launch, the comparison shows the on-demand run instead.
Troubleshooting
The Launch button is disabled.- Hover it — the checklist popover names the unmet gate. A warning icon next to it lists the exact actions (“Assign an AI agent”, “Assign a phone number to this agent”, “Verify the SIP trunk on this agent’s BYO numbers, or assign another number”, …).
- The agent has numbers, but each one has an unresolved payment failure, and a number pending release is never used as a caller ID. The checklist still counts these numbers, so Number assigned to agent can show as met. Resolve the payment issue (see When a Payment Fails) or assign another number to the agent.
- The agent’s only numbers are BYO numbers, including the number you set up with Forward an existing number during onboarding, and none of them is on a Verified SIP trunk. Unlike the payment case above, the checklist doesn’t count these numbers, so Launch stays disabled. Give the agent a number bought in the Studio, or ask your provider to set up and verify a SIP trunk and attach it to the BYO number. See Adding a Number You Already Own.
- Check that the campaign’s agent still has a usable phone number. The dialer reads the agent’s numbers on every dial, so if its last number was unassigned, or every number left is pending release for non-payment or is a BYO number that isn’t on a Verified SIP trunk, the run stays active but places no calls, with nothing on screen saying why. Assign a number bought in the Studio to the agent and dialing resumes from the next dial — no relaunch needed. A BYO number starts dialing again once its SIP trunk is Verified.
- Check the status badge tooltip. Schedule-window pauses resume automatically. For an insufficient-balance pause, a Manage billing shortcut appears in the header: add funds, then select Resume — the campaign doesn’t restart on its own. Resume is refused with “Campaign balance is still insufficient to resume. Add funds and try again.” while the balance is still short.
- With Auto Top-up on and working, the dialer tops up the wallet and keeps dialing. With it off — or on but suspended (the billing page says “Auto top-up is temporarily suspended”) — the run pauses when the balance runs low. A call already in progress ends at the next minute if the balance can’t cover it.
- Your wallet balance is too low to start, and auto top-up is off or suspended. On the Overview tab of Manage → Billing, add funds with Add more credits or select Enable Auto Top-up, then launch again.
- The run couldn’t start when you launched it, and Studio showed “Campaign launch failed. Please retry after some time.” The campaign is editable again. Select Relaunch to try again, which creates the next run. See Campaign Statuses.
- Blocked is a compliance outcome, not an error — open the contact’s dial history to see the reason (for example, missing consent or a number on your internal do-not-call list).
- To fix the data, stop the run (or wait for it to finish), correct or re-upload the list on the Contacts tab, then Relaunch. A run works from its own copy of your contact list, taken at launch, so changing the list never changes contacts that are already Blocked in the current run. Only the next run uses the corrected list. The Contacts tab is locked while a run is in progress, including while it’s paused.
- A contact that is still on your internal do-not-call list, or has opted out, is blocked again. If the suppression is correct, there is nothing to fix.
- Stop the current run first, or Resume it if it is paused.
Where to go next
Compliance & DNC
What the TCPA decisions and block reasons mean.
Call History
Campaign calls also appear in Call History with Caller ID, To, and Direction columns.