> ## Documentation Index
> Fetch the complete documentation index at: https://docs.runspark.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Multi-Line Dialing

> Dial 2, 3, or 5 contacts at once and talk to the first human who answers — with FCC abandoned-call compliance built in

Multi-line mode dials 2, 3, or 5 contacts simultaneously and connects you to the first human who answers. Instead of listening to ringing and voicemail greetings between conversations, you go from one live person to the next — it's the standard way high-volume teams 3x their conversations per hour.

Once you're connected, everything works like a [single-line call](/phone/dialer): live transcription, AI coaching, contact context, dispositions, recording, and AI summaries.

## Starting a Session

<Steps>
  <Step title="Switch to Multi-Line">
    In the Phone tab, switch the dialer from single-line to **Multi-Line** mode.
  </Step>

  <Step title="Choose your parallel count">
    Pick 2x, 3x, or 5x — how many contacts dial at once. The count is fixed for the session; start a new session to change it.
  </Step>

  <Step title="Build your call list">
    Filter your CRM to the contacts you want to work: by stage, lead score, source, or custom fields.
  </Step>

  <Step title="Preview the list">
    Review who's queued before anything dials. Remove anyone who shouldn't be called.
  </Step>

  <Step title="Start the session">
    The first batch dials immediately.
  </Step>
</Steps>

## How a Batch Works

1. **Spark dials N contacts at once** — 2, 3, or 5 depending on your parallel count.
2. **Answering-machine detection sorts each line.** The first human to answer is connected to you instantly. Machines are logged as voicemail. Lines nobody picks up time out as no-answer.
3. **You talk.** The call runs like any single-line call — transcript, coaching, contact panel.
4. **You set a disposition**, and the next batch fires automatically. No clicking between calls; the session keeps feeding you conversations until the list is done or you stop it.

## Choosing Your Parallel Count

Higher parallel counts mean more conversations per hour, but also a higher chance two people answer at once. Match the count to your list:

| Count | Best for                              | Why                                                                       |
| ----- | ------------------------------------- | ------------------------------------------------------------------------- |
| 2x    | Fresh leads and high-value lists      | Answer rates are high — you want almost every pickup to reach you cleanly |
| 3x    | Standard aged leads                   | The balance most teams run day to day                                     |
| 5x    | Deep aged lists with low answer rates | Most lines won't pick up, so dialing wide keeps conversations flowing     |

## The Session Dashboard

While a session runs, the dashboard shows live stats:

* **Connected** — humans you've spoken with
* **Voicemail** — lines that hit answering machines
* **No answer** — lines that timed out
* **Abandoned** — extra simultaneous answers that couldn't be connected to you

You can **pause** and **resume** at any point, or **end the session** early. Sessions survive a page refresh — if your browser reloads mid-session, you pick up where you left off.

<Note>
  **Compliance is built in.** When two humans answer at the same time, you can only take one — the extra call is handled with an identification message rather than dead air. Spark monitors your abandoned-call rate against the FCC's 3% threshold; if you exceed it, your parallel count steps down automatically until the rate recovers.
</Note>

## Tips for Better Sessions

* **Segment before you dial.** A list filtered to one stage and lead type keeps your opener consistent and your dispositions meaningful.
* **Watch the abandoned count.** If it's climbing, your list is answering better than your parallel count assumes — end the session and restart at a lower count.
* **Disposition fast.** The next batch fires when you disposition, so quick, accurate dispositions keep the session moving and your CRM clean.

## FAQ

<AccordionGroup>
  <Accordion title="Can I change the parallel count mid-session?">
    No — the parallel count is fixed per session. End the session and start a new one to dial at a different count.
  </Accordion>

  <Accordion title="What happens if my browser refreshes during a session?">
    Sessions survive a page refresh. When the page reloads, your session — list, progress, and live stats — is right where you left it.
  </Accordion>

  <Accordion title="What counts as an abandoned call?">
    When two humans answer simultaneously, you can only take one. The extra call is handled with an identification message and counted as abandoned. Spark tracks this rate against the FCC's 3% threshold and steps your parallel count down automatically if you exceed it.
  </Accordion>

  <Accordion title="Do voicemails count against my abandoned rate?">
    No. Answering-machine detection logs machines as voicemail dispositions — only extra human answers count as abandoned.
  </Accordion>
</AccordionGroup>

## Related

* [Power Dialer](/phone/dialer) — the in-call experience, coaching, and the full disposition reference
* [Phone Numbers](/phone/phone-numbers) — number reputation matters even more at multi-line volume
* [Inbound Calls & Forwarding](/phone/inbound-calls) — handle the callbacks your sessions generate