Campaigns

A Campaign dials every contact in an audience using a selected voice flow. OliAI handles rate limiting, retries, and concurrency automatically.

Campaign Lifecycle

DRAFT → IN_PROGRESS → COMPLETED
          ↕
        PAUSED
          ↓
       CANCELLED

Creating a Campaign

Intent Labels

Every campaign has a set of intent labels — the outcomes OliAI uses to classify each call (for example Interested in Paying, Callback Requested, Not Reachable). The set is scoped to that one campaign; editing labels on one campaign never affects another.

When you create a campaign, OliAI pre-fills the labels from your Use Case Type. You can then tailor them on the campaign detail page, in the Intent Labels panel.

A Not Reachable label is always present (even on Custom campaigns) because OliAI relies on it for unanswered calls and retry logic.

Configuring labels

In the Intent Labels panel (while the campaign is in DRAFT) you can:

• Add up to 10 custom labels on top of the preset.
• Rename any label, including system labels.
• Pick a colour from the palette — used for the label badge across the app.
• Toggle Positive — marks the label as a positive outcome (used in reporting).
• Set Retry Behaviour — what tagging a call with this label does to retries (see Suppression).
• Reorder labels with the up/down arrows. Order sets the priority OliAI uses to resolve ambiguous calls (higher = preferred).
• Delete a label — except system labels (e.g. Not Reachable, Wrong Number), which are required and cannot be removed.

How Calls Get Tagged

After every call ends, OliAI automatically assigns one intent label — no manual review needed:

• Answered calls — the transcript is sent to your flow's AI model, which picks the single best-matching label from your configured set.
• Unanswered calls (no answer, busy, failed before connecting) — tagged Not Reachable automatically, with no AI cost.
• Too short / no speech — calls with under ~5 seconds of speech, or no recognisable content, are flagged Undetermined for your review.
• AI unavailable — if classification fails after several retries, the call is flagged Undetermined so nothing is silently dropped.

Tagging completes within about a minute of the call ending. Tags are visible in the campaign's Call Log and on the contact's call history. Analysis runs entirely within your own organization's AI configuration — transcripts are never shared across organizations.

Suppression & Do Not Contact

Each intent label has a Retry Behaviour that decides what happens when a call is tagged with it:

When a suppression-flagged intent is tagged, the contact is removed from the retry queue immediately.

Suppressed contacts

Filter the campaign Call Log by Suppressed to see them, with the suppressing intent as the reason. Click Restore on a row to put the contact back in the queue — the action is recorded in the security log.

Do Not Contact registry

Org-wide suppressions appear under Do Not Contact (admin), listing the phone number, source campaign, suppressing intent and date. OliAI checks this registry in real time before every call — first attempt or retry — so a number is never dialled once listed.

Journey Rules

Journey Rules decide what happens next for a contact based on the intent label their call was tagged with — so the right follow-up happens instantly, no matter the volume, without anyone reviewing calls by hand.

In the campaign detail page, the Journey Rules panel shows a row per intent label. For each one, pick a next action:

Multiple labels can use the same action (e.g. both Not Reachable and No Answer → retry in 4 hours).

Editing rules on a running campaign

Unlike intent labels (which lock at launch), Journey Rules can be changed at any time, including while the campaign runs. Changes apply to new tagging events only — contacts already routed are not re-evaluated.

If a contact is retried and their new call gets a different intent label, the rule for that new label takes over — so routing always reflects the latest outcome.

Starting a Campaign

From the Campaigns list or the campaign detail page, click Start Campaign.

Monitoring Progress

Open a running campaign to see the live progress dashboard:

Frequency Cap

The frequency cap limits how often a single contact is called — an org-wide safeguard shared across every campaign, so a contact who appears in several campaigns is protected by one combined limit.

Org Admins configure it in Settings → Call Frequency Cap:

Before every call — first attempt or retry — OliAI checks the contact's recent call count across all your campaigns:

• Daily cap reached → the call is deferred to the next calendar day (in the contact's timezone if known, otherwise the campaign's).
• Weekly cap reached → deferred until the rolling window frees a slot.

Deferred calls re-queue automatically and still respect the campaign's calling window on the new date. Each deferral is recorded on the call with a reason (Frequency Cap - Daily / Weekly), and the campaign dashboard's Deferred — Frequency Cap metric counts them.

Pausing and Resuming

• Pause — Stops initiating new calls. Calls in progress complete normally.
• Resume — Continues from where it left off, dialing remaining contacts.

Viewing Campaign Calls

In the campaign detail page, scroll to the Calls section to see every call attempt:

Click CSV in the Calls toolbar to download an enriched results export — every contact with their intent tag, status, timestamp and attempts.

Cancelling a Campaign

Click Cancel Campaign. This:

• Stops all new dials immediately
• Allows in-progress calls to finish
• Sets the campaign to CANCELLED — it cannot be resumed

Call Audio Playback

For completed calls, click the Play button in the calls table to listen to the recording. Audio is a dual-track recording of both the customer and the AI.

Retry Policy

OliAI automatically retries calls that fail (no answer, busy, network errors). The Retry Policy panel on the campaign detail page lets you tune this per campaign.

The panel shows a plain-English summary of the policy, e.g.:

Up to 3 attempts. Wait 2 hours after attempt 1, 6 hours after attempt 2. Retries only within Mon–Fri 9:00–18:00.

Edge cases

• 0 retries — each contact gets a single attempt. A connected call is marked Completed; an unanswered one is marked Failed (its true outcome).
• Retry Expired — if a retry would fall past the campaign's end date, it's skipped rather than dialed.

Calls are also not retried when a journey rule suppresses them, or when the contact is on Do Not Contact.

Best-time-to-call optimisation

Toggle Best-time-to-call optimisation in the Retry Policy panel to schedule each retry at the time a contact is most likely to answer — always inside the calling window. Off by default.

When on, OliAI picks a preferred hour band per contact, in order:

1. Individual — their own answered-call history (needs ≥ 3 answered calls with some variation in time-of-day).
2. Demographic — if individual data is thin, the answer patterns of similar contacts (same age band / city tier / language), provided the group is large enough.
3. Default — otherwise the org's best-performing time band, or just the normal retry schedule.

The retry is placed in that band on the next valid day within the calling window; if the band never fits the window, the next valid slot is used instead. It still respects the frequency cap and the campaign's retry delay.

Customer-requested callbacks

When a contact asks to be called back at a specific time, OliAI can capture that during the call and automatically schedule the next attempt then.

The campaign call log shows Callback: <time> (with an (adjusted) note when moved) and a Mark fulfilled action to close it out manually — fulfilled callbacks are skipped by the dialer. The scheduled time is also included in the write-back payload as callback_time.

Real-time campaign dashboard

The Campaigns list page is a live operations view across all your campaigns. While any campaign is In Progress or Paused, it auto-refreshes every 30 seconds; a Live · updated Xs ago indicator shows freshness (turning amber with a stale warning if an update is overdue). A manual Refresh button and use-case / status filters are at the top.

An organization summary sits above the list:

Each campaign card adds a strip of real-time metrics:

The health badge flags campaigns that need attention — for example Low connection rate, Few positive outcomes, or Awaiting more results before enough calls have resolved to judge. Hover (or read the line under the metrics) for the reason.