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

# Signals Overview

> Signals are tags applied to calls to classify and segment them, based on data known before the call or insights captured during it.

Signals are tags applied to calls to classify or segment them — each acts as a condition checked on every call. You can segment on pre-call data (Marketing Data) and in-call insights.

Examples:

* **Call and marketing segmentation:** calls from existing customers vs. new leads.
* **Caller insights:** segments from a search ad talking 5 minutes or longer.
* **Conversation insights:** landing pages driving buying-outcome calls.

## Types of Signals

**Pre-call Signals** are evaluated from caller marketing or demographic data, or other information known before the call. This includes any Signal Builder Signal that doesn't rely on Keyword Spotting, and can inform call tracking or routing before the call begins — for example, routing current customers to support.

**In-call Signals** include information from during the call, such as duration and spoken words or phrases. This includes all AI Signals plus Signal Builder Signals that use phrase-spotting or keypress rules, or call duration. These process in real time as the call happens, are available for reporting right after, and can feed other marketing apps — bidding, website content, audiences, suppressing ads, retargeting, and updating your CRM.

You can also apply Signals via API or a manual spreadsheet upload from outside Invoca.

## Creating and managing Signals

Log in and select **Signal** in the gray menu ribbon to see your list of Signals with relevant columns. Create a new Signal as:

* **Rule-based:** define conditions using Signal Builder.
* **AI Signal:** predicted by machine learning from the conversation.
* **Keypress Signal:** for call treatments.

Pause a Signal to stop processing it, or delete it to remove it from your account. Some historical records are kept when a Signal is deleted, but the column is removed from reports. Editing a Signal's name isn't recommended, since it disrupts historical reporting and any integrations that reference it by name.

You can also change a Signal's value on a given call to True, False, or Not Applied to correct inaccuracies — this improves reports, rep evaluation, and Signal AI training. See the guide on manually overriding Signal results for more. Deleting a Signal used in an integration may cause errors in that integration.

### Signal Discovery

Signal Discovery uses unsupervised machine learning to cluster calls into themes and topics, each pre-trained and ready to become a new AI Signal. See the guide on Signal Discovery & Detection for more.

## Utilizing Signal insights

* Send Signal data to other marketing apps via integrations.
* Use Signals in call treatments to customize the caller experience.
* Combine Signals into scorecards.
* Track Signals in Invoca reports and share them with your team.

## Where to go next

* [Using Signal Builder](/s/article/using-signal-builder)
* [Keyword Spotting](/s/article/keyword-spotting)
* [Signal Discovery & Detection](/s/article/signal-discovery-detection)
* [Signal Management: Manual Result Overrides](/s/article/signal-management-manual-result-overrides)
* [Pausing Signals](/s/article/pausing-signals)