Agent Visibility Analytics dashboard in Optimizely Analytics
- Optimizely Analytics
- Optimizely Content Management System (SaaS, or CMS 12 and 13)
The Agent Visibility Analytics dashboard in Optimizely Analytics lets you view AI platform traffic and intent so you can optimize your site's performance for AI. The dashboard is powered by a combination of content delivery network (CDN) log data and Opal-generated insights.
With the data in Agent Visibility Analytics, you can track AI traffic trends and intent, identify the AI platforms visiting your site most often, and discover which webpages are most popular among them. This lets you refine your content for AI optimization and leverage high-performing pages to improve overall engagement, because you understand why AI is crawling your site and the context behind those crawls that Opal can provide.
The dashboard ships with a curated set of charts (provided by Optimizely) that you can copy to a new Analytics dashboard and refine to fit your preferences.
Prerequisites
You can use Agent Visibility Analytics if you meet the following requirements:
- Have an Opti ID account.
- Have access to Optimizely Opal.
- Have access to an Optimizely Content Management System (SaaS, or CMS12 and 13) instance, either headless or coupled.
- Have access to an Optimizely Analytics instance.
- Host your frontend with Optimizely (if using a headless approach).
If you are not a CMS customer or do not meet the requirements but would like to bring your own Cloudflare log data, contact your Optimizely Account Executive to learn more.
Open the Agent Visibility Analytics dashboard
The Agent Visibility Analytics dashboard is the system-level dashboard managed by Optimizely. See the Dashboard overview for more on how dashboards work in Analytics.
To locate the Agent Visibility Analytics dashboard, use the following methods:
- Use the Optimizely Analytics home page filters to display only dashboards.
- Use the search bar to enter keywords and locate dashboards by category, name, or creator.
Click Agent Visibility Analytics.
To make changes, duplicate the dashboard first. Click More (⋮) > Save As. Enter Name, Description, and select the Folder where you want to save the duplicate. Click Apply.
The original system dashboard stays unchanged. Updates Optimizely makes to the system dashboard do not affect your duplicated copy.
You can also add individual tiles from the system dashboard to another dashboard. Hover over a tile, click More (⋮) > Add to another Dashboard. Choose Pick a Dashboard to add the tile to an existing dashboard or New Dashboard to create a dashboard with the tile.
When you have your own copy, you can use Opal to add additional data to it. See Enrich your data with Opal.
View Agent Visibility Analytics
View dashboard charts and intent data
To analyze the dashboard without manually sifting through it, run the Agent Visibility Analytics Insights agent in Opal Chat. The agent examines your dashboard and emails you a structured report of trends, anomalies, and insights. See Agent Visibility Analytics Insights agent.
The Agent Visibility Analytics dashboard is organized into three sections.
Top LLMs Visiting You and What LLMs Want From Your Site
These two sections help explain which large language models (LLMs) visit your site most often and where your content is gaining traction across AI platforms. You can track LLM requests over time to see how changes to your content and strategy influence AI visibility.
- Requests by LLM – Shows the total volume of requests to your site, broken down by AI platform. Use this to see which AI platforms request your site the most or least, and how that visibility changes over time.
- Requests by Top 10 Pages – Shows the top 10 most requested pages by AI platforms. Use this to see which pages or assets have the most visibility with AI platforms and therefore need to be kept up to date most frequently.
- LLM Requests Over Time – Shows how requests from AI platforms change over time. Use this to see how content optimizations or market shifts influence request volumes.
- Page Requests by LLM – Shows the full list of site URLs and requests broken down by each AI platform. Use this to see which pages get the most engagement with one AI platform versus another. If you want to optimize for a specific AI platform, this view shows how its requests to the same page compare to other AI platforms and how its interest varies across content types.
Why Are LLMs Visiting Your Site?
This section explains why LLMs are requesting content from your site. Crawls alone do not reveal what drives the activity. This view shows what does.
Requests are categorized into three intent types based on User Agent values:
- User Action – AI systems retrieving content in real time to answer a user's question. This also indicates which AI platforms most of your userbase prefers to use.
- Search – AI indexing and organizing your content for future retrieval.
- Training – AI collecting content to improve and train models.
Use this section to identify which pages are being used for training versus indexing, and which are actively powering answers delivered to real users. This shows where your content influences AI-generated experiences, and where it goes underused.
Data is available from November 28, 2025 onward.
Filter data with parameter and filter tiles
The Agent Visibility Analytics dashboard includes parameter and filter tiles that you can use to narrow the data displayed in each visualization. For background on the different tile types and how they behave, see Supported tiles.
Use the Time Range parameter tile to change the time window. Select a preset option (such as Last 7 days or Last 30 days), or choose a custom range. All linked visualization tiles on the dashboard refresh automatically when you change the parameter.
Use the Page and LLM filter tiles to narrow results to one or more of your pages or LLMs, respectively. Select the operator and the hostname values you want to include, then apply. For more on how filter conditions and operators work, see Filtering.
If you duplicated the dashboard, you can add additional parameter and filter tiles to your copy. Click + Add Tile, select Parameter or Filter, configure the tile, and link it to the visualizations you want it to control. See Dashboard creation walkthrough for more information.
Click the Tile name to expand the tile.
Segment and create beelines
To gain more granular insights from your data, you can segment individual dashboard tiles. This feature lets you apply filters based on specific cohorts or attributes directly to a single visualization, giving you focused analysis without altering the entire dashboard.
- Right-click the dashboard tile you want to segment.
- Select Segment by Cohort or Segment by Attribute.
- Select an option from the drop-down list to adjust the chart on the tile.
- (Optional) Click Reset to revert the tile to its original state.
For details on the segmentation, see Segment dashboard tiles.
Beelines let you conduct ad-hoc analysis directly from your dashboard explorations. You can drill down into specific data points, explore user behavior, or create new cohorts, transforming a static view into an interactive launchpad for deeper investigation.
To create beelines, complete the following:
- Right-click the exploration from which you want to create a beeline.
- Select a beeline from the following options: Drill down on point, Drill down on series, Explore users, Create cohort from users, and Lookup users.
Use Undo or Redo to reverse actions, or click Reset to revert to the original state.
See Create beelines within a dashboard for more information.
Download data from a tile
You can download any tile on the dashboard as a comma-separated values (CSV) file. Hover over the tile and click More (⋯) > Download As > CSV or Excel. The CSV preserves the column formatting shown in the tile.
Enrich your data with Opal
Looking at request data by URL can be overwhelming, and a raw URL by itself lacks context. With your Analytics instance, you can use an Opal AI agent to add additional context to your dashboard by grouping pages into logical facets that you define.
Facets are custom categories you define and Opal assigns to each URL, such as funnel stage, content topic, or page goal. They let you analyze AI request and intent data by meaningful groupings instead of one raw URL at a time.
For background on how Opal works inside Analytics, see Optimizely Opal Chat for Analytics and Exploration generation with Opal.
For example, you can see how requests differ by:
- Funnel stage (Opal can classify URLs as Top, Middle, or Bottom funnel based on definitions you provide).
- Content topic.
- Page goal.
- Location.
- Any other facet that is meaningful to your content strategy.
Contact Optimizely Expert Services for help creating the facets agent. To create the agent yourself, follow the steps in the following sections.
Define the facets agent
- Open Opal and create an agent dedicated to facet classification for Agent Visibility Analytics.
- Provide the agent with the list of facets you want to track, along with a short description of each facet value (for example, what makes a page "Top funnel" versus "Middle" or "Bottom").
- Point the agent at your site so it can read each URL and classify pages into your defined facets. Instruct the agent to create a CSV file with one row per URL and its classified facet value.
Send facet data to your dashboard
- From the facets agent, publish the classification output as a data source available to Optimizely Analytics. To do this, tell Opal to "Upload this CSV to the APP_NAME app as a dataset called DATASET_NAME". This uses the oa_csv_upload Opal tool. See Organization to find the App name.
- In Analytics, go to Data > Datasets > + New Dataset > Source Dataset. See Source dataset.
- Find the table you asked Opal to create. To preview the selected table, click Preview in the Source Preview section. Click Confirm.
- Enter a Name and optionally a Description.
- Click the Primary Key tab.
- Click + Add Primary Key, select URL from the drop-down list and click Apply.
- Click Save.
- Click the Related Datasets tab.
- Click + Add Related Dataset. Select One to many from the Cardinality drop-down list and Requests dataset from the Target Dataset.
- Select URL from the Current Dataset column and Request URL from the Optimizely-managed Target Dataset column.
- Click Save changes. You only need to configure this once, unless you change your facet definitions later.
- Navigate back to your dashboard. You only need to configure this once, unless you change your facet definitions later.
The facet values become available as catalog properties that you can reference in tiles, parameter values, and filter conditions across the dashboard.
Catalog properties are the fields in Optimizely Analytics (also called entities, properties, or attributes) that describe your users and events, such as "country" or "device type." You use them to group, break down, or filter your data, and you find them through the entity search when building a report or exploration.
Build a tile with facet data
- In your duplicated dashboard, click + Add Tile and select New Visualization. See Explorations for the available exploration types.
- Choose an exploration type (for example, Event Segmentation) and build a measure such as Requests by LLM or LLM Requests Over Time. For more on configuring the chart itself, see Visualization.
- In the Segmentation module, add an attribute backed by your new facet (for example, Funnel stage) so the visualization is broken down by that facet.
- Click Run Exploration, then Add to Dashboard.
- Optionally add a Filter or Parameter tile that lets viewers toggle between facet values, and link it to your new tile.
You can now see request and intent data broken down by the facets that matter to your strategy, rather than by raw URLs.