Profound connector tools

  • Updated

The Profound connector tools in Optimizely Opal let you analyze how AI systems are interacting with your content, providing insights into citations, sentiment, and brand visibility. Complete the following steps to enable them.

Add the Profound Analytics for Opal app in OCP

In the OCP App Directory, complete the following:

  1. Click on Profound Analytics for Opal.

  2. Click Install App. 

Go to Settings > Profound API Configuration.

  1. Enter your Profound API key. See Profound's documentation on Authentication to learn more. 
  2. (Optional) Enter an API Base URL. Entering a URL overrides the Profound API base URL. See Profound's documentation on the Base URL for information.

  3. Click Save API Configuration. 

    Save API Configuration

Configure default category (optional)

After configuring your Profound API Key, you can select a default Profound category to query when using the Profound tools in Opal. If you have multiple categories, you can set a default here and query others in your Opal prompt, or you can leave this section blank, which may result in Opal asking which category to query.

To configure a default category, complete the following:

  1. Expand the Default Category Configuration section.
  2. Select the Default Category drop-down list.
  3. Select the default category or search to find it.
  4. Click Save Category Configuration.

Register the Profound connector tool in OCP

After you Add the Profound Analytics for Opal app in OCP and optionally Configure default category, complete the following in OCP:

  1. Click Settings > Add to Opal.

  2. Click Add to Opal for the Opal Account you want to add the Profound connector tools to. Click Remove from Opal to remove the connection. 

You can now use the Profound connector tools within Opal. 

Profound connector tools

After you add and register the Profound connector tools, you can call the following tools in Opal. Click on a tool to expand it to learn when to use the tool, available parameters, and example prompts. If you do not provide a required parameter, Opal prompts you for it.

raw_answers – Provides access to unprocessed prompt and answer data for custom analysis and reporting, including prompts, mentions, citations, themes, and metadata.
  • When to use
    • For custom analysis and dashboards – If you need the granular, unprocessed data to build your own dashboards or perform specialized analysis that is not covered by the aggregated reports.
    • To understand the full context of AI interactions – If you want to see the exact prompts given to AI systems, the mentions, and citations they generated, and the themes identified.
    • For in-depth research – If you are conducting research into AI behavior and need the foundational data.
  • Parameters
    • start_date – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • (Optional) category_id – The Category ID from the get_categories tool, for example, your_category_id.
    • (Optional) filters – A JSON array of filters, for example, [{"field":"prompt_type", "operator":"is", "value":"visibility"}].
      • Available fields for filtering include the following – 
        • prompt_type
        • model
        • region
        • topic
        • theme
      • Available operators include the following – 
      • is
      • not_is
      • in
      • not_in
      • contains
      • not_contains
      • contains_case_insensitive
      • not_contains_case_insensitive
      • matches
    • limit – The number of records to return. The default is 50.
    • offset – The starting position for pagination. The default is 0.
  • Example prompts
    • Get all raw visibility data for the 'Technology' category (ID: tech_123) from January 1st to January 7th, 2023.
    • I need raw data for all prompts related to the 'AI Model X' from October 1st to October 31st, 2025, limited to 100 results.
    • Retrieve all raw answers for the 'Marketing' category (ID: marketing_456) for the entire month of November 2025, focusing on prompts that contain the word 'campaign'
get_domains – Lists domains available for log analysis, which are required parameters for the agent_analytics_bots tool and the agent_analytics_logs tool.
  • When to use
    • To understand your monitored websites – If you want to see a list of all the websites that Profound is tracking for AI system interactions and traffic.
    • Before using agent analytics tools – If you need a domain name for the agent_analytics_bots tool or the agent_analytics_logs tool and you do not know which domains you configured for monitoring.
  • Parameters
    • None
  • Example prompts
    • What domains are available for analysis?
    • List all the websites I can monitor with Profound.
    • Show me the domains being tracked for bot and traffic analytics.
agent_analytics_bots – Provides insights into bot citations and training data usage to monitor how AI systems access and reference your content.
  • When to use
    • To see which AI bots are citing your content – If you want to know which AI assistants are referencing your website pages and how often.
    • To understand which pages are being used for AI training data – If you want to identify which parts of your website are most frequently scraped by AI training bots.
    • To monitor bot traffic patterns – To analyze the methods, status codes, and paths bots are accessing on your domain.
    • To identify specific bot activity – If you are interested in the behavior of bots from a particular provider or of a certain type.
  • Parameters
    • domain – Your website domain (for example, www.example.com). Use the Get domains tool if you do not know which domains are available.
    • start_date – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • metrics – What you want to measure. Only "count" is available.
    • dimensions – How you want to break down the data (for example, path, bot_name, bot_types).
    • (Optional) filters – A JSON array of conditions to narrow down your data (for example, [{"field":"bot_types", "operator":"contains", "value":"ai_assistant"}]).
    • (Optional) limit – The maximum number of records to return. The default is 100.
    • (Optional) offset – The starting position for pagination. The default is 0.
    • (Optional) order_by – A JSON object for sorting the results (for example, {"count":"desc"}).
  • Example prompts
    • Show me how many times AI assistants cited my blog pages on www.example.com between January 1st and January 31st, 2024, broken down by bot name and page path.
    • Which 10 pages on mywebsite.com were most frequently accessed by AI training bots last month (October 2025)?
    • Give me a count of all bot accesses to yourcompany.net in the last week (October 23rd to October 30th, 2025), showing the status code and method used.
agent_analytics_logs – Provides insights into website traffic and log analysis to monitor how users and systems access your content.
  • When to use
    • To understand overall website traffic – If you want to see the total number of requests to your site over a period.
    • To analyze user and bot access patterns – If you need to break down traffic by method (for example, GET or POST), status code (for example, 200 or 404), or user agent.
    • To identify popular pages or resources – If you want to see which paths on your website are most frequently accessed.
    • To troubleshoot issues – If you need to investigate specific status codes or methods that might indicate problems.
    • To monitor specific user agents – If you want to track traffic from a particular browser, device, or known bot.
  • Parameters
    • domain – Your website domain, for example, www.example.com. Use the Get domains tool if you do not know which domains are available.
    • start_date  – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • metrics – What you want to measure. Currently, only "count" is available.
    • dimensions – How you want to break down the data. You can choose from the following:
      • timestamp
      • method
      • host
      • path
      • status_code
      • ip
      • user_agent
      • referer
      • bytes_sent
      • duration_ms
      • query_params
    • (Optional) filters – A JSON array of conditions to narrow down your data, for example, [{"field":"method", "operator":"is", "value":"GET"}].
      • Available fields for filtering includes the following: 
        • path
        • method
        • status_code
        • user_agent
      • Available operators include the following:
        • is
        • not_is
        • in
        • not_in
        • contains
        • not_contains
        • contains_case_insensitive
        • not_contains_case_insensitive
        • matches
    • (Optional) limit – The maximum number of records to return. The default is 100.
    • (Optional) offset – The starting position for pagination. The default is 0.
    • (Optional) order_by – A JSON object for sorting the results, for example, {"count":"desc"}.
  • Example prompts
    • Show me the total number of successful GET requests to www.mywebsite.com last week (October 23rd to October 30th, 2025).
    • List the top five most accessed paths on example.org in September 2025, along with the number of times each was accessed.
    • How many requests did blog.company.com receive from mobile user agents in the last 30 days (October 1st to October 30th, 2025), and what were the different status codes returned?
get_categories – Retrieves available category IDs and names for your account. These categories organize and filter data by business domains or market segments. Use the get_categories tool to get the required category_id parameter for the citation_report tool, the sentiment_report tool, and the visibility_report tool.
  • When to use
    • Before using other Profound tools – If you need a category_id for tools like citation_report, visibility_report, or sentiment_report, and you do not know the available categories or their IDs.
    • To understand your account's categorization – If you want to see how your content is organized within Profound.
    • To discover available analysis scopes – If you are looking for different business domains or market segments you can analyze.
  • Parameters
    • None
  • Example prompts
    • What categories are available in Profound?
    • List all the content categories I can use for reporting.
    • Can you show me the IDs and names of all Profound categories?
citation_report – Provides insights into how AI systems reference and cite content across the web, including metrics of mentions, references, and citation patterns.
  • When to use
    • To identify authoritative sources – If you want to find out which domains or pages are most frequently cited in a specific category.
    • To understand market share and competitive positioning – If you want to analyze how often your company or specific topics are cited compared to competitors.
    • To track content influence – If you want to see how AI systems are referencing your content over time.
    • To analyze citation patterns – If you need to break down citations by region, topic, or AI model.
  • Parameters
    • category_id – The ID of the category you want to analyze. Use the get_categories tool to get a list of available categories.
    • start_date – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • metrics – What you want to measure. Available values are "count" (total number of citations) and "share_of_voice" (your share of citations within the category).
    • dimensions – How you want to break down the data. Available values include the following:
      • hostname
      • path
      • date
      • region
      • topic
      • model
      • tag
    • (Optional) filters – A JSON array of conditions to narrow down your data, for example, [{"field":"hostname", "operator":"is", "value":"example.com"}].
      • Available fields for filtering include the following:
        • hostname
        • path
        • region
        • topic
        • model
        • tag
      • Available operators include the following:
        • is
        • not_is
        • in
        • not_in
        • contains
        • not_contains
        • contains_case_insensitive
        • not_contains_case_insensitive
        • matches
    • (Optional) limit – The maximum number of records to return. The default is 10.
    • (Optional) offset – The starting position for pagination. The default is 0.
    • (Optional) order_by – A JSON object for sorting the results, for example, {"count":"desc"}.
  • Example prompts
    • For the 'Technology' category (ID: tech_123), show me the top 5 hostnames with the most citations in the last month (October 2025).
    • I want to see the share of voice for different topics within the 'Digital Marketing' category (ID: digital_marketing_456) for the last quarter (July 1st to September 30th, 2025). Please exclude any citations from 'spamdomain.com'.
    • Can you give me a daily count of citations for my website mycompany.com in the 'Ecommerce' category (ID: ecommerce_789) for the first two weeks of November 2025?
sentiment_report – Provides insights into sentiment data and emotional responses across companies and topics.
  • When to use
    • To understand company reputation – If you want to identify companies with the most positive or negative sentiment.
    • To analyze sentiment across themes – If you need to see which topics generate positive or negative sentiment for your company or competitors.
    • To track sentiment over time – If you want to monitor changes in sentiment for specific entities or topics over various date intervals.
    • To identify areas for improvement – If you need to pinpoint specific areas that are causing negative sentiment.
  • Parameters
    • category_id – The ID of the category you want to analyze. Use the get_categories tool to get a list of available categories.
    • start_date – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • metrics – What you want to measure. Available values are include the following:
      • positive
      • negative
      • neutral
      • total_mentions
    • dimensions – How you want to break down the data. Available values are include the following:
      • hostname
      • path
      • date
      • region
      • topic
      • model
      • tag
      • asset_name
      • theme
    • (Optional) date_interval – How to group the dates – day, week, month. The default is day.
    • (Optional) filters – A JSON array of filters, for example, [{"field":"hostname", "operator":"is", "value":"example.com"}].
      • Available fields for filtering include the following:
        • hostname
        • path
        • region
        • topic
        • model
        • tag
        • asset_name
        • theme
      • Available operators include the following:  
        • is
        • not_is
        • in
        • not_in
        • contains
        • not_contains
        • contains_case_insensitive
        • not_contains_case_insensitive
        • matches
    • (Optional) limit – The number of records to return. The default is 10.
    • (Optional) offset – The starting position for pagination. The default is 0.
    • (Optional) order_by – A JSON object for sorting the results, for example, {"positive":"desc"}.
  • Example prompts
    • For the 'Financial Services' category (ID: financial_123), show me the top 5 companies with the most positive sentiment last quarter (July 1st to September 30th, 2025).
    • I want to analyze the sentiment across different topics for 'MyCompany' in the 'Retail' category (ID: retail_456) for the month of October 2025. Show me the topics with the highest positive sentiment.
    • What was the daily negative sentiment trend for 'Product X' in the 'Consumer Goods' category (ID: consumer_789) during the first two weeks of November 2025?
visibility_report – Provides insights into company visibility metrics and performance data within specific categories. It helps you track how prominently your company or specific assets display in AI-generated content.
  • When to use
    • To identify market leaders – If you want to see which companies have the highest visibility scores in a particular category.
    • To benchmark against competitors – If you need to compare your company's visibility performance against top performers.
    • To monitor daily performance changes – If you want to track how your visibility score changes over time with daily, weekly, or monthly granularity.
    • To analyze visibility by topic or region – If you need to understand how your visibility varies across different topics or geographic regions.
    • To identify content gaps – If you want to see which of your assets are contributing most to your visibility, and where there might be opportunities to improve.
  • Parameters
    • category_id – The ID of the category you want to analyze. Use the get_categories tool to get a list of available categories.
    • start_date – The beginning of the date range for your analysis in YYYY-MM-DD format.
    • end_date – The end of the date range for your analysis in YYYY-MM-DD format.
    • metrics – What you want to measure. Available values are count (total mentions) and share_of_voice (your share of mentions within the category).
    • dimensions – How you want to break down the data. Available values include the following:
      • hostname
      • path
      • date
      • region
      • topic
      • model
      • tag
      • asset_name
    • (Optional) date_interval – How to group the dates. date_interval is only used with the date dimension. The default is day. Available options include the following:
      • day
      • week
      • month
      • year 
    • (Optional) filters – A JSON array of filters, for example, [{"field":"hostname", "operator":"is", "value":"example.com"}].
      • Available fields for filtering include the following:
        • hostname
        • path
        • region
        • topic
        • model
        • tag
        • asset_name
      • Available operators include the following:
        • is
        • not_is
        • in
        • not_in
        • contains
        • not_contains
        • contains_case_insensitive
        • not_contains_case_insensitive
        • matches
    • (Optional) limit – The number of records to return. The default is 10.
    • (Optional) offset – The starting position for pagination. The default is 0.
    • (Optional) order_by – A JSON object for sorting the results, for example, {"count":"desc"}.
  • Example prompts
    • For the 'Software' category (ID: software_123), show me the top five companies with the highest visibility scores between January 1st and January 31st, 2024.
    • Track the daily visibility performance for 'MyCompany' in the 'Cloud Computing' category (ID: cloud_456) for the month of October 2025.
    • What is the share of voice for different regions in the 'Fintech' category (ID: fintech_789) for the last quarter (July 1st to September 30th, 2025)?

If you use Opti ID, administrators can turn off generative AI in the Opti ID Admin Center. See Turn generative AI off across Optimizely applications.