> For the complete documentation index, see [llms.txt](https://docs.dstillery.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.dstillery.com/data-and-audience-products/audiences/custom-ai-audiences.md).

# Custom AI audiences

Build a predictive audience from your first-party customer and conversion signals.

### What it is

Custom AI Audiences solve the problem of underutilized first-party data. Instead of using your CRM list or pixel data for retargeting only, Dstillery's multimodal AI takes those signals and builds a predictive model that finds your next best customers across the entire programmatic landscape.

The model learns from your actual customers, not generic demographic proxies. It scores 400M+ devices daily to find people who look and behave like them.

Your seed can include site visitors, conversion pixels, e-mail lists, or campaign logs. Use DS-1 to [create a pixel](/using-ds-1/agents/create-a-pixel.md) that collects site-visitor and conversion signals. These methods support DS-1 self-service or managed service.

Email and CRM lists require managed service through your Dstillery Client Services team.

### Why it's different

{% columns %}
{% column %}

#### Before

Your first-party data, whether it is a CRM list, pixel, or conversion log, is typically limited to retargeting or basic lookalike modeling. These approaches rely on single signals and do not scale well for finding new customers.
{% endcolumn %}

{% column %}

#### After

Dstillery combines all of your data into one predictive model. It uses real behavioral patterns to find new customers at scale, not just people who resemble them on paper.
{% endcolumn %}
{% endcolumns %}

### How it works

{% stepper %}
{% step %}

#### Prepare the seed

First-party seed data can include site visitors, conversion pixels, or campaign logs. These seeds support DS-1 self-service or managed service. For a DS-1 workflow, set up a [Create a pixel](/using-ds-1/agents/create-a-pixel.md). It collects site-visitor and conversion signals.

Client Services can manage pixel setup and seed preparation. Email and CRM lists require Client Services for matching and ingestion. Dstillery requires at least 1,000 seeds to build a predictive model.
{% endstep %}

{% step %}

#### Analyze digital journeys

Dstillery's multimodal AI analyzes full digital journey patterns across 400M+ devices.
{% endstep %}

{% step %}

#### Receive and refresh

A custom predictive model is built and delivered to your DSP within 24 hours. The model refreshes every 24 hours, so your audience never goes stale.
{% endstep %}

{% step %}

#### Activate everywhere

One model powers every activation method, from user segments to custom bidding algorithms.
{% endstep %}
{% endstepper %}

### Key benefits

{% hint style="success" %}

* Built from your own data, so the model reflects your actual customers, not generic segments.
* Activatable as ID-based or cookieless via ID-free technology, flexible across all environments.
* Uses event-level behavioral modeling across customers' full digital journey, not basic overlap matching.
  {% endhint %}

### Good fit if you

* Have first-party data to work with. Set up site visitors, conversion pixels, and campaign logs in DS-1 or via managed service. Use Client Services for email and CRM lists.
* Are running prospecting campaigns and want to scale beyond retargeting.
* Want your custom audience built from real behavioral patterns, not a generic lookalike.

### Pricing

See [Pricing](/data-and-audience-products/audiences/pricing.md) for current rates.