Channel Automation · Calling
Contact Center · Calling · Agent Command Center · Field Coach

Calling, as a complete package

A dialer that remembers every call — so retries pace themselves, no two agents call the same person, callbacks happen when the customer asked, no single number burns out, and every attempt shows up in the numbers a manager and a coach both read.

Why this exists

Every one of these is a real thing a client asked for in the same call: a dialer that knows who was called, when, from which number, and what happened — and acts on it. Before this, the platform placed calls well and forgot them immediately.

Before

  • Two agents on one queue both call the same lead this morning.
  • "Already called today" lives in one browser for 12 hours — switch laptops, the leads come back.
  • Retries happen when someone remembers. Callback times live on sticky notes.
  • Every call to a contact goes out from the same number, ten times if needed — the fastest way to a spam label.
  • 1,671 leads in a queue, agents only ever see the same 300.
  • Nobody can say which attempt number books appointments, or how many dials each number made today.
  • Coaching (Field Coach) and dialing (Calling) don't know about each other.

After

  • A lead being dialed is hidden from everyone else for 10 minutes. Universal, no setup.
  • The memory is on the server, per contact — the same on every machine.
  • A queue's cadence decides when a lead comes back; a callback returns at the time the customer named.
  • A daily limit per number and a no-repeat rule, one switch per workspace.
  • The counter says 1/300 · 1,671 in queue and refills — one session works the whole list.
  • Team Stats shows outcomes by attempt and dials per number vs its limit.
  • Field Coach sees which attempt a call was, samples first impressions, and is one click from Team Stats.

The package — six pieces, one memory

Everything below reads and writes the same record: this contact was dialed N times, last from this number, last outcome this, eligible again at this time. That shared memory is what makes the pieces complement each other instead of being six features.

1 Call Queues + Cadence

  • A live dialing list by lead status, age, or a field, assigned to agents
  • Cadence: how many attempts, how long between, what stops it
  • Badge: ready · cooling · done
Admin sets · Agent works

2 Call Guide & Power Dialer

  • One lead at a time with context, or auto-dial through a list
  • Attempt chip: Attempt 3 of 6
  • Honest counter, refills past the served slice
Agent

3 Callbacks

  • Choose Callback → enter the time the customer asked for
  • Lead returns at that time; listed in the Callbacks tab
Agent

4 Number rotation

  • Daily limit per number, no-repeat to the same person
  • One switch per workspace, off by default
  • Texting numbers never change
Admin

5 Inbound numbers

  • Label each line by campaign; agents see it on the incoming call
  • Routing rules can ring specific agents for a line
Admin sets · Agent sees

6 Team Stats + Field Coach

  • Dials, connects, reached, appointments — by agent, by attempt, by number
  • Coaching link → Field Coach Team; Field Coach samples by attempt
Manager

Cadence, explained

A cadence is the retry schedule a queue follows on its own. Three settings: how many times to try a lead, how long to wait after each try, and which outcomes end it. Once set, an agent never decides "should I call this person again today?" — the queue simply doesn't show the lead until it's due, and stops showing it once it's done.

Play it through. Dial Luigi, pick what happened, and watch the week fill in. The sentence under the calendar says what the queue is doing and why.

Queue cadence6 attempts · 4h → 1d → 3d → 7d · busy 30m
Attempt0 of 6
Luigi is ready — he's never been dialed, so he sits at the top of the queue (never-dialed leads first).
What each setting means
SettingPlain English
Max attempts"Give up after 6 tries." The lead is done for this queue.
Wait after attempt 1 / 2 / 3 / 4+"After the first miss, hide them 4 hours. After the second, a day…" The last row repeats.
Busy retry"Busy means they're on the phone — try again in 30 minutes instead."
Stops the cadence"Booked, do-not-call, or transferred? Never dial again from this queue."
Priority"Serve brand-new leads before due retries" (speed-to-lead) — or the reverse.
Turn "Cadence on" off in the simulator to see today's behaviour: the lead is hidden only in your own browser, for the rest of your shift, then comes straight back — no pacing, no stop.

Two things every queue gets, cadence or not

  • A lead being dialed right now is hidden from other agents for 10 minutes.
  • Callback offers a time field; the Callbacks tab lists it.

A "done" lead never comes back on its own

  • That is deliberate — a do-not-call must never be re-dialed by accident.
  • An owner or admin can reset a contact's cadence; it starts again at attempt one.

An agent's shift, in eight steps

Click through. Each step shows the screen the agent is on and what the package is doing underneath.

  1. 1 · Open Calling, pick a queue
  2. 2 · Read the lead
  3. 3 · Start call
  4. 4 · Save the outcome
  5. 5 · A callback with a time
  6. 6 · Next lead — nobody else has it
  7. 7 · End of the slice — it refills
  8. 8 · Your day, in numbers
The one habit that makes all of this work: save an outcome on every call. An outcome is what starts the wait, honours the callback time, and stops the cadence. A ring-out is filled in for you from the carrier result; an answered call needs your click.

Set up & monitor — the admin's side

Set up once

DoWhereWhy
Build a queue and assign agentsAgent Command Center → Call Queues → New QueueLead statuses, contact age, one custom field — the same list the agent gets, counted live
Switch on a cadenceQueue card → CadenceA suggested table pre-fills (6 · 4h → 1d → 3d → 7d · busy 30m); edit and it saves
Turn on number rotation (optional)Settings → Contact Center → Number rotationDaily limit + no-repeat. Off by default; texting numbers never change. If you use Caller ID Intelligence it still picks first
Label inbound lines (optional)Settings → Contact Center → Inbound numbersAgents see the campaign on the call; rules can ring specific agents for a line
Route a line to agents (optional)Agent Command Center → Access & Routing → Call Routing Rules"Number the caller dialed" works even for unknown callers

Monitor every morning

What the buckets mean

Ready

  • Can be served to an agent right now
  • Never-dialed first, unless you chose due-first

Cooling

  • Dialed; waiting for its next attempt
  • Or being dialed this minute (10-minute hold)

Done

  • Booked, do-not-call, transferred, or out of attempts
  • Only an owner/admin brings it back

The coaching loop — Calling and Field Coach together

Calling knows how a call happened — which attempt, which number, what outcome. Field Coach judges how it went — score, objections, coaching. They now hand each other context without merging into one number.

Attempt context on every recording

  • Recent Calls and the Field Coach call header show First attempt / Attempt 4
  • A coach reads a sixth-touch call differently from a first touch

Smarter sampling

  • Auto-analyze can pick first attempts that connected — the first impression
  • Or reached, ran out of attempts, never booked — the richest coaching set

One click, not two systems

  • Team Stats → Coaching link per rep → Field Coach Team
  • Contact rate on one screen, coaching score on the next
Field Coach never moves the cadence. Its follow-up drafts stay drafts and its next-action is advice. Only a human's outcome or the carrier's result decides when a lead is dialed again.

What it lets us measure that we couldn't before

By attempt
Dials, connects, people reached, and appointments for attempt 1, 2, 3, 4+ — the first honest answer to "how many touches does it take?"
Per number, today
Dials against the daily limit for every caller ID — the same counter the dial path reads, so the card and the calls can't disagree
Who chose the number
How many dials went out on a rotated number, a caller-ID-intelligence pick, or the legacy default
Cooling / done
How much of a queue is paced vs exhausted — is the list being worked or worn out?
Callbacks kept
Callback outcomes with a time, and whether the lead was served at it
Coverage
Share of dials with a saved outcome — the discipline number that decides whether contact rate may be shown at all

All of these come from one definition — the same function feeds the Team Stats tab and the daily calling email, so a manager comparing the two never finds a gap.

What it does not claim

It does not manage number reputation or caller-ID branding

Scanning carriers, removing spam labels, and branded caller ID are a vendor's job (Caller ID Intelligence, where a workspace uses it). Rotation spreads load and avoids repeats; it doesn't score reputation. If the vendor is on, it picks first — rotation acts only when it doesn't.

It only counts dials made inside the platform

A team still dialing from another tool has attempts the ledger never sees, so an attempt count can read lower than the rep's own notes. The fix is to dial from here.

A contact rate is withheld when it would mislead

Below 10 dials or 70% coverage, Team Stats shows a dash and the reason — not a confident percentage. That's the platform's standing reporting rule, applied here too.

The daily limit is a setting, not our finding

Twelve is a starting point; carriers are commonly reported to flag numbers above the low teens per day. Agree the right value with whoever manages your reputation.

Posting a subsource to an outside CRM is not wired yet

Inbound labels and subsource codes are stored on every call so that step can be added; today the label reaches the agent and the routing rules.

Start in three steps

  1. Pick one queue (New Leads is the usual) and switch its Cadence on with the suggested table. Watch ready · cooling · done for a week.
  2. Tell agents one thing: save an outcome on every call, and enter the time when a customer asks for a callback.
  3. After the first week, read Team Stats → By attempt and shorten or lengthen the cadence from what actually books. Then decide on number rotation with your reputation vendor.
Nothing to switch on globally. It's already available to every workspace, and every piece is inert until an admin configures it.