Profound: Setup guide

Profound: Setup guide#

This guide explains how to collect data from Profound.

Introduction#

Profound tracks how brands appear in AI-generated search results across major large language models (LLMs) such as ChatGPT, Perplexity, and Gemini. It measures brand visibility, share of voice, and sentiment in AI-generated responses to search queries, enabling enterprise marketing and competitive intelligence teams to benchmark their brand’s presence in AI search.

Limitations#

Collecting data from Profound comes with the following limitations:

  • Data for the Sentiment report type is only available up to 8 July 2026. For later dates, use the Sentiment (v2 beta) report type.

Prerequisites#

Before you start collecting data from Profound, perform all of the following actions:

  • Ensure you have a Profound account with API access enabled.

Creating a datastream to collect data from Profound#

The basics of creating a datastream to collect data from any data source are explained in our guide to Collecting data in Adverity. This guide contains information about the specific steps to create a datastream to fetch data from Profound.

Authorization: Allow Adverity to access Profound#

Before creating an authorization to give Adverity access to Profound, perform the following actions:

  • Contact Profound’s support team to obtain your API Key.

To set up a new authorization to fetch data from Profound, follow these steps:

  1. In the Authorization step of the datastream wizard, click Access Profound using your credentials.

  2. Click Next.

  3. In API Key, enter your Profound API Key.

  4. Click Authorize.

As a result, Adverity has access to Profound using the credentials you provided.

Configuration: Choose the data you want to collect from Profound#

To choose what data to collect and customize the Profound datastream configuration, follow these steps:

  1. (Optional) Rename your datastream.

  2. In Report type, select the type of report to collect. The available report types are:

    • Sentiment (default) — brand sentiment expressed by LLMs in response to queries

    • Sentiment (v2 beta) — an updated version of the Sentiment report

    • Visibility — how often your brand appears in AI-generated results

    • Citations — which pages and domains are cited by LLMs when mentioning your brand

    • Citations (v2 beta) — an updated version of the Citations report

  3. For the Sentiment (v2 beta) report type, in Assets, select the assets for which to collect data. If you leave this option empty, data for all assets is collected.

  4. In Categories, select the product or brand categories for which to collect data. You must select at least one category. This option isn’t available for the Sentiment (v2 beta) report type, which instead uses the category of each selected asset.

  5. For the Citations (v2 beta) report type, in Entity, select the entity type for which to collect data: Domain (default), Page, or Citation category.

  6. In Metrics, select the metrics to include in the report. Available metrics differ by report type.

  7. In Dimensions, select the dimensions by which to break down the data. Available dimensions differ by report type.

  8. For the Citations (v2 beta) report type, if Entity is set to Domain, in Scope, select whether to include all cited domains (All, default) or only your owned domains (Owned).

  9. For the Sentiment (v2 beta) report type, if Dimensions includes Theme or Claim, select Include cited websites to also return the websites contributing to each theme or claim.

  10. In Date interval, select the time granularity for the report.

  11. (Optional) In Regions, select the regions for which to collect data. If you do not select any regions, data for all regions is collected.

  12. (Optional) In Models, select the LLMs for which to collect data. If you do not select any models, data for all models is collected.

  13. (Optional) In Category topics, select specific topics within the selected categories.

  14. For the Citations (v2 beta) report type, in Citation categories, select the citation categories to filter by. If you leave this option empty, all categories are included.

As a result, Adverity fetches data from Profound according to the configuration you set.

What’s next?#

Contents