Skip to main content

Open in AthenaHQ

app.athenahq.ai/responses

​
Purpose

The Responses page serves as the central hub for reviewing exactly what AI models are saying about your tracked topics and prompts. It aggregates every individual response collected by AthenaHQ across major AI engines (like ChatGPT, Claude, Gemini, and Perplexity) so you can read the raw output and understand the context behind your metrics. Customers use this page to evaluate how their brand is mentioned, which competitors are appearing alongside them, and what sources the AI models are citing. By exploring the actual text of these responses, marketers can identify gaps in their content, discover new competitors, and verify the accuracy of AI outputs.

​
What’s on the page

​
Page Header

  • Title: Displays “Responses”.
  • Back Button: An optional left-pointing arrow appears if the user navigated here from a specific context (like clicking through from the Sankey visualization), allowing them to return to the previous view.

​
Filter Bar & Saved Views

  • Saved Views: A dropdown menu to save, load, rename, or delete combinations of filters for quick access later.
  • Filter Chips: A row of active filters. Users can refine the table using a wide variety of criteria (see the full list below).
  • Add Filter / [+]: Opens a menu to add new filters to the bar.
  • Clear All: Appears when non-default filters are active. Resets the view to the default state.
  • Columns Button: Opens a popover to toggle the visibility of table columns and reorder them via drag-and-drop.
  • Download Button: Opens the Export Responses dialog to download a CSV.

​
Responses Table

The core of the page is a paginated, sortable, and resizable spreadsheet displaying all AI responses matching the current filters. Columns in the table:
  • Date: The UTC date the response was collected (formatted MM/DD/YY).
  • Model: The AI model that generated the response (e.g., ChatGPT, Claude, Gemini, Perplexity, AI Overview, Copilot, AI Mode, Grok, DeepSeek). Displays as a colored badge with the model’s icon.
  • Mentioned: Indicates if your tracked brand was mentioned. Shows “Yes” (indigo badge) or “No” (orange badge if competitors were mentioned, neutral gray otherwise).
  • Competitors Mentioned: Displays logos of up to 4 competitors mentioned in the response. If more than 4 appear, a “+N” badge is shown. Hovering over a logo reveals the competitor’s name.
  • Sources: Shows “Yes” or “No” indicating if the response cited any sources.
  • Cited: Hidden by default. Shows “Yes” or “No” indicating if your tracked website’s domain was explicitly cited as a source.
  • Attributed Citation: Hidden by default. Shows “Yes” or “No” indicating if the response contains an attributed citation linking back to the tracked website.
  • Market Position: Hidden by default. The numeric rank of your brand among all companies mentioned (1 = mentioned first).
  • Sentiment: Hidden by default. The sentiment of the response toward your brand (Positive, Neutral, or Negative).
  • Prompt Type: Indicates whether the prompt is “Branded” (high-intent) or “Non-Branded” (discovery), shown as a badge.
  • Country: A flag icon representing the geography associated with the prompt or response.
  • Location: Location tag(s) associated with the prompt. Only visible if the website’s subscription plan includes location-based tracking.
  • Personas: Included in the column list for filtering purposes, but renders no visible content in the table cell by default.
  • Prompt: The text of the prompt that triggered the response, truncated. Matches from active searches are highlighted.
  • Prompt Variation: The specific variation of the base prompt used for this run.
  • Attributes: Hidden by default. Brand attributes detected in the response. Positive attributes are green; negative are red.
  • Response: A truncated plain-text preview of the actual AI response. Matches from active searches are highlighted.
  • Response Count: Displays the current viewing range (e.g., “Showing 40 of 1,234 responses”).
  • Load More: Fetches the next set of responses to append to the table.

​
Available Filters

Users can filter the table using the following options:
  • Date Range: Restricts responses to those collected within a specific time window.
  • Response Search: Searches the actual text of the AI responses.
  • Source Search: Filters for responses that cite a specific domain.
  • Models: Filters by the AI models that generated the response.
  • Competitors: Filters for responses mentioning selected competitors (supports “is any of” and “is not any of”).
  • Source Mentions Competitor: Filters for responses where a cited source mentions the selected competitor(s).
  • Prompt Tags: Filters by tags applied to the prompts.
  • Personas: Filters for responses associated with selected personas.
  • Locations: Filters for responses associated with selected locations.
  • Topics & Prompts: Filters for specific topics or individual prompts.
  • Prompt Status: Filters for active, paused, or all prompts.
  • Sentiment: Filters for responses with Positive, Neutral, or Negative sentiment.
  • Brand Mentioned: Filters for whether the brand was or was not mentioned.
  • Source Mentions Brand: Filters for whether a cited source does or does not mention the brand.
  • Has Sources: Filters for responses that do or do not cite sources.
  • Has Attributed Citation: Filters for responses that do or do not have an attributed citation.
  • Prompt Variation: Filters for base prompts only, fan-out variations only, or both.
  • Brand Rank (Market Position): Filters by the numeric rank of the brand in the response.
  • Attributes: Filters for responses containing specific brand attributes.
  • Brand Identifiers: Filters for responses where specific brand identifier strings were matched.
  • Has Shopping Products: Filters for responses containing shopping product results.
  • Has AI Overview: Filters for responses that include a Google AI Overview.
  • Has Organic Results: Filters for responses containing organic search results.
  • Has People Also Ask: Filters for responses containing a ‘People Also Ask’ section.
  • Has Related Searches: Filters for responses containing related searches.
  • Organic Results Domain: Filters responses where the organic results contain a specific domain.

​
Drilldowns

  • Response Detail Drawer: Triggered by clicking any row in the table. This slide-in panel shows full details: date, model, prompt type, mention status, market position, sentiment, competitors mentioned (with logos), and sources cited (with favicon pills). It also displays the full AI response text, complete with inline citation pills and highlighted brand/competitor mentions.
  • Edit Response Dialog: Triggered by clicking the pencil icon in the drawer header. A modal where users can manually override and correct the “Mentioned” status, Market Position (rank), Sentiment, and the order of competitor mentions for that response.
  • Flag Response Dialog: Triggered by clicking the flag icon in the drawer header. A modal where users can report data issues with the response to the AthenaHQ team.
  • Export Responses Dialog: Triggered by clicking the Download icon. A modal for configuring a CSV export.
  • External URLs: Clicking any citation pill inside the response text or any source badge in the drawer opens the cited website in a new tab.

​
What you can do here

  • Add Filter: Click the + Filter button to open the field picker and add new filter criteria.
  • Clear All: Click the orange “Clear All” button to reset non-default filters.
  • Columns: Click the “Columns” button to open the popover, check/uncheck columns to toggle visibility, and drag handles to reorder them.
  • Download: Click the download icon button to open the Export Responses dialog.
  • Load More: Click the “Load More” button at the bottom of the table to load additional rows.
  • Export (in Export Responses dialog): Click to download a CSV of the filtered responses.
  • Cancel (in Export Responses dialog): Closes the export modal without downloading.
  • Adjust Export Quantity: Inside the Export dialog, use the + and - buttons to increment/decrement the export size by 1,000, or click Reset to 1,000 or Export all.
  • Quick-select Export Quantity: Click the 1,000, 5,000, or 10,000 buttons in the Export dialog to rapidly set the limit.
  • Table row click: Click anywhere on a response row to open the Response Detail drawer.
  • Edit (pencil icon): In the Response Detail drawer, click the edit icon to open the Edit Response dialog and correct data.
  • Flag (flag icon): In the Response Detail drawer, click the flag icon to report an issue with the response.
  • Navigate Up / Navigate Down: In the Response Detail drawer, click the up/down arrows (or use keyboard Arrow Up / Arrow Down) to switch to the previous/next response without closing the drawer.
  • Close drawer: Click the X button in the top right of the drawer to close it.
  • Save View: In the Saved Views dropdown, save the current filter combination as a reusable preset.
  • Load View: In the Saved Views dropdown, apply a previously saved preset.
  • Delete View (Destructive): In the Saved Views dropdown, permanently delete a saved preset.
  • Back (arrow): Click the back arrow in the page header to return to the referring page (e.g., if you arrived here from a specific data visualization).
  • Mobile menu actions: On small screens, tap the ⋯ menu to access the Columns toggle (submenu with checkboxes) and the Download button.

​
Data shown

The table displays actual AI model responses collected by running your configured prompts against engines like ChatGPT, Claude, Gemini, Perplexity, Copilot, AI Mode, Grok, DeepSeek, and Google AI Overviews. Mention statuses are determined by matching the text against your configured brand identifiers, brand attributes, and competitor list. Competitor logos are pulled dynamically from AthenaHQ’s logo storage.

​
Common workflows

1. Review AI responses for a specific prompt
  1. Navigate to the Responses page.
  2. Open the “Topics & Prompts” filter and select the desired prompt.
  3. Browse the filtered response rows to see what different AI models said.
  4. Click a row to open the Response Detail drawer.
  5. Read the full AI response text with brand and competitor highlights.
2. Export filtered responses to CSV
  1. Apply desired filters (e.g., a specific Date Range, Model, or Brand Mentioned = Yes).
  2. Click the Download icon button in the filter bar.
  3. In the Export Responses dialog, adjust the quantity using the +/- buttons or quick-select presets.
  4. Click “Export”.
  5. Wait for the progress toast to complete; the CSV file will download automatically.
3. Correct inaccurate response data
  1. Find the response in the table and click its row to open the drawer.
  2. Click the pencil (Edit) icon in the drawer header.
  3. In the Edit Response dialog, update the Mentioned status, Market Position, Sentiment, or competitor ranks.
  4. Click “Save Changes”.
  5. The drawer and table will update to reflect the corrected data.
4. Investigate which sources are cited alongside brand mentions
  1. Apply the “Brand Mentioned = Yes” filter.
  2. Optionally, add a “Source Search” filter for a specific domain.
  3. Click a response row to open the drawer.
  4. Scroll to the Sources section to see all cited domains with their favicon badges.
  5. Click a source badge to open the source URL in a new tab.
5. Customize the table columns for a focused view
  1. Click the “Columns” button in the filter bar.
  2. In the Columns popover, check or uncheck columns (e.g., enable Market Position, Sentiment, or Attributed Citation).
  3. Drag column rows to reorder them.
  4. Close the popover; the table updates immediately and your preferences are saved to your browser.

​
Empty, loading, and error states

  • Empty: If filters yield no results, the table body displays the message “No responses found.”
  • Loading: An animated spinner (AnimatedLoader) is shown centered in the table area while the initial data fetch is in progress.
  • Errors: There is no explicit error UI in the table itself, as the system automatically retries failed requests behind the scenes. If a CSV export fails, a red toast notification will appear in the corner of the screen.
Linked from:
  • Prompts page: Clicking a prompt row drills through to this page, pre-filtered to that prompt.
  • Olympus / Analytics page: Clicking “View Responses” links pre-filter by date range and model.
  • Sources / Sankey visualization: Clicking a Sankey node navigates here with Sankey-derived response IDs and the source domain pre-applied.
  • Direct URLs: Arriving with ?attributedCitation=true in the URL pre-applies the “Attributed Citation = Yes” filter.
  • Sidebar navigation: The main “Responses” nav item.
  • Tour system: The Prompts guided tour automatically navigates to this page.
Links to:
  • External source URLs: Cited sources open in a new tab directly from the detail drawer.
  • Referring page: The back button in the header navigates to the page specified in the ?returnUrl query parameter (e.g., back to the Sankey visualization).

​
Common support questions

Why can’t I see the Location column? The Location column (and location-based filters) only appear if your subscription plan has the location-based tracking feature enabled. Why are some columns missing from my table? To keep the view uncluttered, columns like Market Position, Sentiment, Attributed Citation, Cited, Personas, and Attributes are hidden by default. Click the “Columns” button in the filter bar to check the boxes next to the columns you want to see. I clicked a dot in the visualizer and came to this page, but I don’t see the filter for it? When you click through from the Sankey visualization, the system applies a “hidden” filter for those specific response IDs on the server side. Because this is a highly specific query, it doesn’t display as a normal chip in the filter bar, and you cannot remove it manually. Switching to a different website will clear this session state and redirect you back to the visualizer. Why is the “Load More” button grayed out? The “Load More” button becomes disabled when you have reached the end of the available data (i.e., when the number of displayed responses equals the total count of responses matching your filters).