The Dialing Mode Matrix — Seven Pacing Rules, One Table

Dialing modes compared: manual to agent-assisted, in one matrix

Every outbound dialer answers the same question — when does the next call go out, and how many at once — and every dialing mode is a different answer. This page is the comparison hub: the full matrix in one table, then the matchups people actually search — power dialer vs predictive dialer, preview vs predictive, single-line vs multi-line — plus the assisted-dialer category. Each mode links to its own deep-dive page; for the umbrella taxonomy and the robocall distinction, start with the auto dialer guide.

The dialing mode matrix

Seven modes cover the territory, from full human control at the top to full delegation at the bottom. Definitions follow taxonomies published by Genesys Cloud, Five9, NICE, Convoso, and Koncert as of July 2026 — vendors genuinely disagree on labels (more below).

One warning before the matchups: the labels are not standardized. In Genesys Cloud’s documentation, “power” mode dials multiple numbers per idle agent while “progressive” is strictly one call per agent; Five9 defines power dialing as a fixed calls-per-agent ratio; Convoso’s power mode is a fixed, non-adaptive dial level; and Koncert brands its single-line product an “AI Power Dialer.” NICE defines a progressive dialer as one call per agent — and also says it presents contact information before dialing, which most vendors call preview. Same words, different machines: compare mechanics, not mode names.

Power dialer vs predictive dialer

This is the matchup that decides most purchases, and the difference is who the pacing serves.

A power dialer — in the sequential, one-line sense most sales teams mean — serves the individual agent. It dials one contact, the agent hears the ring live, and when the call ends the software advances to the next record. Every answered call has a human already on the line; throughput is capped by the list’s answer rate, but nothing is abandoned. When a vendor’s “power” mode dials multiple lines per agent, as Genesys and Five9 define it, it behaves more like the parallel row above — the strict one-call-per-agent version is documented under the name progressive dialer.

A predictive dialer serves the team. It launches calls before any agent is free, pacing against live statistics — NICE’s glossary lists connect rates, average talk time, agent breaks, and abandon rates as the algorithm’s inputs — and, per Convoso’s documentation, disconnects unsuccessful calls before an agent ever hears them, so agents move from one live answer to the next. The cost is the abandoned call: a person answers and no agent is ready. That tradeoff is regulated — FCC rules cap abandoned telemarketing calls at 3% of calls answered live by a person, per campaign over each 30-day period (47 CFR 64.1200, as of July 2026) — and the TCPA autodialer rules guide covers what that means in practice.

Team size settles most of these decisions. Predictive pacing is statistics, and vendors publish their own minimums: Five9 positions predictive for 10+ active agents and progressive for fewer, while Genesys recommends roughly 15 agents for predictive accuracy. Below those ranges the algorithm paces lumpily, which is why small teams usually get more from sequential power or parallel dialing.

Preview dialer vs predictive dialer

These two are opposite ends of the control spectrum, and the comparison is really about who decides.

A preview dialer shows the agent the contact record — name, history, notes — before any call is placed, and the agent chooses to dial or skip, per Genesys Cloud’s definition; Five9’s version delivers the record with CRM history attached. It’s the slowest mode on the board by design: on lists where one conversation carries real weight, the minute spent reading the record is worth more than another dial.

A predictive dialer inverts every one of those choices: the system decides what to dial and when, the agent sees the record as the live call arrives, and skipping isn’t part of the workflow. The practical rule — if your agents would benefit from knowing who they’re calling before the phone rings, preview; if the list is too long for that to matter, predictive.

Single-line vs multi-line dialing

Cutting across the modes above is a simpler axis: how many lines per agent.

A single line dialer runs one call at a time per agent — click-to-dial, preview, progressive, and sequential power dialing all share this shape. The agent is attached before the phone rings: no connection delay for the person answering, no abandoned calls, undivided attention on every dial.

A multi-line dialer — the parallel dialer, in sales-tech vocabulary — dials several numbers simultaneously for one agent and connects the first person who picks up. Orum states the mechanic cleanly: when two people answer, it detects who answered first down to the millisecond, connects the rep to that person, and cancels the other call. Parallel dialing exists because most cold-list dials go unanswered; the extra lines convert dead ringing time into conversations, at the cost of a connection seam the answer-rates deep dive examines honestly.

On Enzo the two aren’t separate products: Starter seats dial 5 lines per agent, Standard up to 14, and lines pool across agents — two Standard agents can work one list as a pooled 28-line session. Enterprise removes the line cap; teams run 30+ lines per agent as needed.

What is an assisted dialer?

An assisted dialer isn’t a pacing rule — it’s a delegation model layered on one. The category splits in two.

Human-assisted. Koncert’s Agent-Assisted Dialer uses what it calls “trained, professional human dialing agents. No robots” — a team that dials on the rep’s behalf and stops dialing the moment the rep is in a conversation. Koncert claims its dialing agents average 125 calls per hour for a rep; that’s the vendor’s own figure, not an audited one.

Software-assisted. The same only-live-conversations outcome, delivered by detection instead of people. Koncert’s single-line dialers automatically filter busy tones, fax tones, bad numbers, phone trees, and voicemails before connecting the rep, and Twilio’s engineering docs describe the pattern underneath most of these products: answering machine detection (AMD) classifies whether a human or a machine answered, and only invokes the agent on a human. AMD has real tradeoffs — the answering machine detection explainer covers how it works and where it misfires. Stated plainly: Enzo does not use AMD; every call an Enzo campaign connects is heard live by your agent from the first hello. The AI end of this category, where parallel dialers add AI filtering and coaching, gets its own comparison in AI parallel dialer vs power dialer.

How to choose a dialing mode

Match the mode to the list and the team, not to a vendor’s philosophy:

• Short, warm, or high-stakes list — preview. The record read before the dial wins the conversation.
• CRM-centered day, modest volume — click-to-dial keeps the call inside the record the rep is working.
• Small team, decent answer rate — sequential power (progressive): steady pace, live person on every connect.
• Cold lists where most dials go unanswered — parallel/multi-line, where conversations-per-hour gains actually live.
• Large floor, long list, short repeatable calls — predictive, sized to your vendor’s published agent minimums.
• Reps who should only touch live conversations — the assisted category, human or software, priced accordingly.
• Compliance-conservative operations — fully manual options exist for this: Five9’s Manual Touch Mode disables automated dialing entirely, and Genesys documents an external-calling mode where the agent places each call by hand.

Most teams don’t live in one row — the same operation runs preview on Tuesday’s past-client list and multi-line on Wednesday’s cold pull. That’s the argument for buying modes as settings, not separate products.

Where Enzo fits in the matrix

Enzo covers the middle of the table — power, predictive, and preview dialing, in single-line and multi-line configurations — on every plan, switchable per campaign. There’s no mode named progressive on the platform; Enzo’s single-line power mode is the same one-call-per-agent machine that name describes. There’s no click-to-dial mode, no AMD, no prerecorded voice, and no AI voice: every mode connects a live agent.

Around the modes sit the things dialing volume actually strains: 35 managed caller IDs per seat on Starter, 100 on Standard, rotated or swapped when reputation dips; all inbound and outbound minutes included; whisper, barge-in, and optional call recording. Pricing is published at /pricing: from $99 per seat per month billed annually, $120 month-to-month, no seat minimum, with the $250 white-glove buildout waived on annual.

Not sure which row your list belongs in? Book a free discovery call — 20 minutes, and if the right mode is one Enzo doesn’t run, we’ll say so.

Mode definitions and mechanics cited from Genesys Cloud, Five9, NICE, Convoso, Twilio, Koncert, and Orum public documentation and product pages, and the abandonment rule from 47 CFR 64.1200 (eCFR), as of July 2026. Vendor terminology and features change — verify with each vendor before you buy.