Skip to main content

Open in AthenaHQ

app.athenahq.ai/responses

​
Purpose

The Responses page is the raw data engine of the Athena platform. It shows every single output collected from AI models across all of the prompts you are tracking. This page allows you to review exactly what the AI engines are saying, verify if your brand or competitors are mentioned, check the sentiment of the responses, and investigate which sources the AI relies on. Customers use this page to drill down into the details of their AI search performance, moving from high-level metrics into the actual text the AI generates to better understand the nuances of their visibility.

​
What’s on the page

Page Header Displays the title “Responses”. If you navigated here from a specific drilldown (like a prompt details page or a visualization), a back-navigation arrow appears next to the title to return you to your previous view. Filter Bar A highly customizable row of filters allowing you to slice your response data. It includes a search box, a date range selector, an “Add Filter” button for applying additional criteria, a Saved Views dropdown, a Columns toggle, and a Download (export) button. Responses Table A spreadsheet-style, infinite-scrolling table listing every AI response that matches your filters. Table Columns:
  • Date: The UTC date the response was collected (formatted as MM/DD/YY).
  • Model: The AI engine that generated the response (e.g., ChatGPT, Claude, Gemini, Perplexity, AI Overview, Copilot, AI Mode, Grok, DeepSeek). Shows as a colored badge with an icon.
  • Mentioned: Whether your tracked brand was mentioned in this response. Shows as a “Yes” or “No” badge. (If “No” but competitors were mentioned, the badge is orange).
  • Competitors Mentioned: Logos of competitors mentioned in the response. Up to 4 logos are visible; if more are mentioned, a tooltip labeled +N appears showing the rest.
  • Sources: Whether the response included any cited sources (Yes/No).
  • Cited: (Hidden by default) Whether your tracked website’s domain appears as a cited source in this response.
  • Attributed Citation: (Hidden by default) Whether the response contains a direct, attributed link back to your tracked website.
  • Market Position: (Hidden by default) The numeric rank of your brand among all companies mentioned in the response (e.g., 1 means you were mentioned first).
  • Sentiment: (Hidden by default) The AI’s sentiment toward your brand in this response (Positive, Neutral, or Negative).
  • Prompt Type: Whether the prompt is Branded (high-intent) or Non-Branded (discovery).
  • Prompt: The prompt text that triggered the response. If you used the search bar, matching terms are highlighted.
  • Prompt Variation: The specific variation of the prompt used (e.g., base prompt or fan-out).
  • Personas: (Hidden by default) Personas associated with this response.
  • Attributes: (Hidden by default) Brand attributes detected in the response, shown as green (positive) or red (negative) badges.
  • Country: A flag icon representing the geography associated with the response or prompt.
  • Location: Specific location tags for the response. (Note: Only visible if the location offering is enabled on your plan).
  • Response: A truncated, plain-text preview of the actual AI response. Matches from the search bar are highlighted.
Row Count Footer Located at the bottom of the table, this text displays “Showing X of Y responses” alongside a manual Load More button (though scrolling down will also automatically load more rows). Response Detail Drawer Clicking any row in the table slides out a detailed panel on the right side of the screen. This drawer shows the complete text of the AI response (with brand and competitor names highlighted), a list of all cited sources, and metadata (Date, Model, Mentioned status, etc.). It also contains buttons to edit or flag the data, and external links to open the cited sources in a new browser tab.

​
What you can do here

  • View response details: Click any row in the table to open the Response Detail Drawer and read the full AI output.
  • Navigate between responses: While the detail drawer is open, click the Up arrow (Previous response) or Down arrow (Next response) in the drawer header to move through the list without closing the panel.
  • Edit response data: In the detail drawer, click the pencil icon (Edit response). This opens a modal where you can manually correct whether the brand was mentioned, the market position, the sentiment, and the list of competitors mentioned.
  • Flag a response: In the detail drawer, click the flag icon (Report issues with this response) to report data quality issues to the Athena team.
  • Copy response ID: While the detail drawer is open, hold the Alt or Option key on your keyboard to reveal the internal Response ID in the header; click it to copy it to your clipboard.
  • Toggle columns: Click the Columns (View Options) button in the top right of the table toolbar to open a checklist. You can check/uncheck columns to show or hide them, and drag them up and down in the list to reorder them in the table.
  • Resize columns: Hover between column headers in the table and drag the handle left or right to adjust width. Double-click the handle to auto-fit the column to its content.
  • Sort data: Click any sortable column header (such as Date, Model, Mentioned, Competitors Mentioned, Market Position, Sentiment, Prompt Type, or Prompt Variation) to sort the table.
  • Apply Filters: Click Add Filter + to restrict the table using a massive variety of criteria. You can filter by Models, Competitors, Prompt Tags, Personas, Locations, Brand Identifiers, Sentiment, and visibility flags like “Has AI Overview”, “Has Shopping Products”, or “Has Attributed Citation”.
  • Search text: Type in the Search box to find responses where the prompt or the AI’s output contains your search string. You can also add specific “Source Search” or “Persona Search” filters via the Add Filter menu.
  • Clear all filters: If you have applied custom filters, a Clear All button appears next to the Add Filter button to instantly reset the table.
  • Save and load views: Click the Saved Views dropdown on the left side of the filter bar to save your current combination of filters for future use, or to load, rename, or delete an existing saved view.
  • Export to CSV: Click the Download (downward arrow) icon in the top right. This opens an Export dialog where you can set how many rows to download (using +/- buttons, typing a number, or using the 1,000/5,000/10,000 quick-select presets) and trigger the CSV export. (Note: This feature requires a paid subscription).
  • Mobile actions: On mobile devices, the filter bar is collapsed. Click Filters to open a mobile drawer of filter options. To access the Columns toggles and the Download button, tap the ... (more) menu.

​
Data shown

The data on this page represents the actual responses generated by AI models when Athena automatically runs your tracked prompts on a set schedule. The competitors, brand identifiers, personas, and locations shown are based on the configurations established in your workspace settings.

​
Common workflows

1. Review AI responses for a specific prompt
  1. Navigate to the Responses page (or arrive automatically by clicking a prompt drilldown on the Prompts page).
  2. If you didn’t arrive via a drilldown, use the Topics & Prompts filter to scope the table to the prompt you care about.
  3. Scan the table, particularly looking for rows where Mentioned is “No” but Competitors Mentioned contains logos.
  4. Click a row to open the Response Detail Drawer and read exactly what the AI said.
  5. Use the Up/Down arrows in the drawer header to quickly review the adjacent responses.
2. Correct inaccurate response data
  1. Locate a response in the table that seems to have miscategorized your brand or a competitor, and click it to open the drawer.
  2. Click the pencil (Edit) icon at the top of the drawer.
  3. In the modal, correct the Brand Mentioned toggle, Market Position, Sentiment, or modify the list of competitors.
  4. Click Save Changes. The table and drawer will immediately refresh to reflect your manual corrections.
3. Export responses to CSV
  1. Apply your desired filters (e.g., a specific date range, a specific model, and negative sentiment).
  2. Click the Download icon button on the far right of the filter bar.
  3. In the Export Responses dialog, adjust the row limit. You can use the + or - buttons, the quick-select presets, or click “Export all” to download the entire filtered dataset.
  4. Click Export. A toast notification will track the progress and confirm when the download completes.
4. Investigate a Sankey visualization click-through
  1. On the Sources > Visualization (Sankey) page, click a specific node or flow.
  2. You will be automatically redirected to the Responses page.
  3. The page will apply a hidden filter to show only the exact responses that made up that Sankey node, and it will highlight the relevant source domain in the table.
  4. Review the filtered rows and click into them to understand context.
5. Filter and search responses
  1. Open the Date Range filter to restrict the timeframe.
  2. Click “Add Filter” to layer on criteria like Model is ChatGPT, Sentiment is Positive, and Has Sources is Yes.
  3. Type a keyword into the Search box to find specific phrases in the AI output.
  4. Once you have a valuable combination, click the Saved Views dropdown and save it so your team can easily access it later.

​
Empty, loading, and error states

  • Loading: While data is being fetched, a spinning animated loader appears in the center of the table area. The filter bar may show gray skeleton placeholders while dynamic dropdown options (like competitor lists) load.
  • Empty: If your filters are too restrictive and return no data, the table body will display a simple “No responses found.” message.
  • Error: If the platform fails to fetch the data, it will automatically attempt to retry silently in the background. If you attempt an action (like flagging or editing a response) and it fails, a red error toast notification will appear.
Linked from:
  • Sidebar Navigation: “Responses” is available as a top-level item in the main menu.
  • Prompts page: Clicking a row or clicking from a prompt’s analytics drawer drills down to this page, automatically filtering for that prompt.
  • Sources > Visualization: Clicking a node on the Sankey chart routes here to show the underlying responses.
  • Olympus Dashboard: Clicking an “Attributed Citation” stat card routes here and automatically filters for responses with attributed citations.
  • Guided Tours: The product tour on the Prompts page includes a “Go to Responses” button that navigates here.
Links to:
  • Previous Page: The back arrow in the header returns you to your previous location (e.g., the Prompts page or the Sankey chart) if you arrived via a drilldown.
  • External URLs: Links in the Response Detail Drawer’s “Sources” section open the cited third-party webpages in a new browser tab.

​
Common support questions

Why can’t I see the Location column in my table? The Location column is only visible if your organization’s subscription plan includes the location-based tracking offering. I know the AI cited us, but why don’t I see the “Cited” or “Attributed Citation” columns? To keep the table clean, these columns (along with Market Position, Sentiment, Personas, and Attributes) are hidden by default. You can easily reveal them by clicking the “Columns” (sliders) icon in the top right and checking their boxes. Why does the export button ask me to upgrade my plan? Downloading raw response data as a CSV is a premium feature gated behind paid subscription plans. Free-plan users will see an upgrade dialog when they attempt to use the export button. I clicked a node on the Sankey chart and was brought here, but I don’t see any filters applied? When you arrive from the Sankey visualization, the platform applies a hidden “Response ID” filter to perfectly match the data you clicked. This filter isn’t visible in the standard filter bar chips to prevent accidental removal, but it is active. Switching websites or navigating away will clear it. I changed my filters, but the URL still says ?promptId=.... Am I stuck viewing one prompt? When you drill down from the Prompts page, the URL parameter forces the table to only show that prompt, overriding your filter bar temporarily. If you want to view all prompts again, either clear the filter, or use the back arrow to return to the Prompts page.