Prompt Universe Run Detail

​

Purpose

The Prompt Universe Run Detail page is where a customer reviews the results of a single Prompt Universe analysis. Prompt Universe is Athena’s prompt-discovery tool: it scans the customer’s keyword landscape, competitor keywords, and social discussions (Reddit, YouTube) to surface AI-search prompts that the customer’s brand is not yet monitoring. The detail page is where the customer reads through those discovered prompts, filters them, understands why each was suggested, and decides which ones to add to their tracked prompt library. This page sits at the end of the Prompt Universe workflow. A customer kicks off an analysis from the Prompt Universe tab on the Prompts page, waits for it to run, and then lands here to inspect what was found. Discovered prompts are grouped by topic so the customer can scan by theme rather than wading through a flat list. Each prompt comes with metadata explaining where it came from (Competitor Gap, Social Signal, Keyword Opportunity, Integrated Signal), where it sits in the buyer journey (TOFU/MOFU/BOFU), what the searcher wants (Informational/Commercial/Transactional), how specific it is (Head/Mid tail/Long tail), and its estimated monthly volume and dollar value. The core action on this page is selecting promising prompts and starting to track them. Once a prompt is tracked, it joins the customer’s regular monitoring library and Athena begins running it against the AI models on the customer’s schedule. The page also exposes lightweight management: renaming the run, switching between recent runs, exporting to CSV, and (for admins) deleting runs.

​

What’s on the page

​

Page header

• Back arrow: returns to the Prompt Universe tab of the Prompts page.
• Run name: the name of the analysis run. Clicking it (or the pencil icon that appears on hover) turns it into an inline editor so the customer can rename the run. Enter saves; Escape cancels.
• Start tracking (N) button: only appears in the top-right when one or more prompts are selected. Clicking it begins the tracking flow. Disabled for non-admins with the tooltip “You need admin access to perform this action”.

​

Run selector

A pill-shaped dropdown in the toolbar showing the current run’s name, a status icon (spinner for in-progress, alert for failed, X for canceled, nothing for completed), and the run’s creation date. Opens a list of up to 4 recent runs for the current website. Admins see a trash icon next to each run in the dropdown to delete it.

​

Toolbar

• Filters dropdown: opens a menu with a “High impact” quick toggle plus sub-menus for Funnel stage, Search intent, Keyword type, and Source. A small numeric badge shows how many filters are active.
• Visualization toggle: opens or closes the visualization panel below the toolbar.
• Search box: free-text search across prompt text, topic name, and recommendation reason.
• Export CSV (download icon): downloads a CSV of all suggestion prompts. The filename is auto-generated as AthenaHQ_<sitename>_universe_<runname>_<date>.csv.

​

Visualization panel

A collapsible card containing three small horizontal bar charts:

• Funnel distribution: counts of prompts by TOFU / MOFU / BOFU. Color-coded sky / amber / emerald.
• Intent distribution: counts by Informational / Commercial / Transactional / Navigational. Color-coded blue / purple / emerald / zinc.
• Volume by topic: top 5 topics by combined monthly search volume.

These charts reflect the current filter and search state (but not the High impact / Added quick filters).

​

Prompts table

The main content. Prompts are grouped into collapsible topic rows. Each topic row shows the topic name, prompt count, and combined monthly volume. Expanding a topic reveals one row per prompt sorted by relevance (high → medium → low). Columns (all sortable via the header chevrons):

• (checkbox): selects the row for batch tracking.
• Prompt: the full prompt text. A “High impact” pill appears on prompts the system flagged as high relevance. An “Added” pill appears once the prompt has been added to tracking.
• Category: how the prompt was discovered. Tooltip: “How this prompt was discovered: Competitor Gap (keywords your competitors rank for), Social Signal (from Reddit/YouTube discussions), Keyword Opportunity (relevant brand or competitor keyword not yet monitored), or Integrated Signal (found in both keyword data and social discussions).”
• Funnel: buyer-journey stage. Tooltip: “Where in the buyer journey: TOFU (awareness: learning about the topic), MOFU (consideration: comparing options), BOFU (decision: ready to buy or act).”
• Intent: what the searcher wants. Tooltip: “What the searcher wants: Informational (learn something), Commercial (compare options), Transactional (take action or buy).”
• Type: keyword specificity. Tooltip: “Search specificity: Head (broad, 1–2 words, high volume), Mid tail (moderate, 2–3 words), Long tail (specific, 4+ words, lower volume but higher intent).”
• Source, which pipeline produced the prompt. Tooltip: “Which pipeline generated this prompt: Keywords (from keyword intelligence analysis) or Social (from Reddit/YouTube discussions).”
• Volume: estimated monthly searches. Tooltip: “Monthly search volume for this prompt. Hover for the exact number.” Volumes under 100 display as “100”.
• Value: estimated monthly traffic value in USD. Tooltip: “Estimated monthly traffic value in USD based on search volume and CPC.”

​

Already tracked section

A collapsed row at the bottom of the table (only visible when the quick filters are off). Tooltip on its info icon: “These prompts are already covered by your current monitoring. They matched existing prompts with high similarity (85%+), meaning the same search intent is already being captured.” Expanding shows the duplicate prompt text in a muted style: these can’t be selected because they’re already in the customer’s library.

​

Footer / summary bar

A thin row beneath the table when the analysis is complete:

• X prompts in Y topics: counts reflecting the current filter and search state.
• Total volume: sum of monthly search volume for the filtered prompts.
• Added N/100 meter: a 12-segment bar showing how many prompts have been moved into tracking out of a fixed reference of 100. Tooltip: “Prompts derived from real market demand signals and already saved in your tracking library for ongoing monitoring.”

​

Confirmation dialog

A modal that appears when the customer clicks “Start tracking”. It reports how many of the selected prompts are duplicates and will be skipped, with Confirm and Cancel buttons.

​

What you can do here

• Go back to the Prompt Universe list: click the back arrow next to the run name.
• Rename the analysis run: click the run name or the pencil that appears on hover, type a new name, press Enter to save. On save failure a toast reads “Failed to rename analysis”.
• Switch to another recent run: open the run selector dropdown and pick a different run (most recent 4 only).
• Delete a run (admins only): open the run selector dropdown, click the trash icon on the row to delete. Soft-deletes the run.
• Apply filters: open the Filters dropdown:
  • Toggle High impact to show only high-relevance prompts.
  • Drill into the Funnel stage, Search intent, Keyword type, or Source sub-menus and pick a value. A check mark indicates the active value; selecting it again clears it.
• Search prompts: type in the search box. Matches against prompt text, topic name, and recommendation reason.
• Show or hide the visualization charts: click the Visualization button in the toolbar.
• Sort the table: click any column header. Chevrons indicate sort direction (descending first, then ascending, then off).
• Expand or collapse all topics: click the chevron toggle in the Topics column header.
• Expand a single topic: click the topic row.
• Select prompts:
  • Click an individual prompt’s checkbox.
  • Click a topic row’s checkbox to select every prompt in that topic.
  • Click the header checkbox to select every visible prompt (across all expanded and unexpanded topics matching the current filters).
• Start tracking selected prompts (admins only): click “Start tracking (N)” in the page header. A confirmation dialog will tell you how many duplicates will be skipped. Click Confirm to add them. On success a toast confirms how many were added and how many were skipped.
• Export CSV: click the download icon in the toolbar. Downloads every suggestion prompt with full metadata (Prompt, Topic, Category, Relevance, Funnel stage, Search intent, Keyword type, Source, Monthly volume, Value USD). Toast on completion: “Exported N prompts to CSV”.
• Run the first analysis (empty state, admins only), if no runs exist, click “Run first analysis” on the empty state.
• View the “Already tracked” section: click the row to expand and see which prompts in the analysis were already covered by existing monitoring.

​

Data shown

Everything on this page comes from a single Prompt Universe analysis run, which is a one-time scan over the customer’s website, competitors, and social signals. The discovered prompts are stored as suggestions; the customer chooses which ones graduate into their actual tracked-prompts library.

• Suggested prompts: the AI-discovered prompts with their categorization, funnel/intent/type/source metadata, and relevance score.
• Search volume and value: estimated monthly searches and USD traffic value, derived from keyword intelligence data and CPC.
• Already-tracked overlap: prompts in the run that match existing tracked prompts at ≥85% similarity, surfaced so the customer doesn’t try to add duplicates.
• Recent runs list: for the run selector dropdown.
• Live duplicate check, when the customer selects prompts and clicks Start tracking, the page checks each selected prompt’s exact text against the website’s current tracked prompts to count duplicates before confirming.

​

Common workflows

Browse and filter discovered prompts

1. Open the page for a completed analysis run.
2. Open the Filters dropdown and toggle High impact, or pick a Funnel stage / Intent / Type / Source.
3. Type into the search box to narrow by keyword or topic name.
4. Click any topic row to expand it and read the individual prompts.
5. Open the Visualization panel to see how the filtered prompts break down by funnel and intent.

Select and track prompts

1. Expand the topics you’re interested in (or use the chevron toggle in the header to expand all).
2. Check individual prompt checkboxes, or click a topic row’s checkbox to take everything in that topic, or click the header checkbox to select every visible prompt.
3. Click “Start tracking (N)” in the page header.
4. Review the duplicate count in the confirmation dialog.
5. Click Confirm. The chosen prompts will appear in the customer’s tracked-prompts library and start running on schedule.

Export the run to CSV

1. Click the download icon in the toolbar.
2. The browser downloads a CSV named AthenaHQ_<site>_universe_<run>_<date>.csv. Note: the export contains every suggestion prompt regardless of the current filters or search.

Switch between recent runs

1. Click the run selector dropdown in the toolbar.
2. Pick one of the 4 most recent runs. The page reloads with that run’s results.

Rename a run

1. Click the run name in the page header.
2. Type the new name in the inline input.
3. Press Enter (or click out) to save.

​

Empty, loading, and error states

• No runs for this website: full-page empty state titled “Discover your prompt universe” with a description of what Prompt Universe does. Admins see a “Run first analysis” button.
• Run not found, if the URL points to a missing or deleted run, the page shows “Analysis not found”.
• Loading: a skeleton placeholder for the header, two stat cards, and a 10-row table, shown while run data fetches.
• Analysis still running: instead of the table, an animated Athena logo with rotating status messages cycles through phases like “Discovering prompts…”, “Analyzing competitor gaps…”, “Scanning social signals…”, “Mapping keyword opportunities…”, “Evaluating search intent…”, “Classifying funnel stages…”, “Scoring relevance…”, “Grouping by topic…”. Filtering and tracking are unavailable until the run completes.
• Run failed / terminated / canceled: surfaced in the run selector via a status icon (red alert for Failed/Terminated, gray X for Canceled).
• Errors: most errors surface as toasts: “Failed to rename analysis”, “Failed to start tracking”, “No prompts to export”, etc.

​

Linked from / links to

• Linked from: Prompt Universe list / tab on the Prompts page (/prompts?tab=prompt-universe): each run row links to its detail page.
• Links to: The Back arrow returns to /prompts?tab=prompt-universe. Tracked prompts created from this page show up in the customer’s main tracked-prompts library.

​

Common support questions

“Why are some prompts marked ‘Already tracked’?” Those prompts matched existing prompts in your monitoring library with 85%+ similarity, so the same search intent is already being captured. You can expand the “Already tracked” row to see them. They can’t be selected here because they’d just be duplicates. “What’s the difference between Competitor Gap, Social Signal, Keyword Opportunity, and Integrated Signal in the Category column?”

• Competitor Gap: keywords your competitors rank for.
• Social Signal: discovered from Reddit/YouTube discussions.
• Keyword Opportunity: a relevant brand or competitor keyword you’re not yet monitoring.
• Integrated Signal: found in both keyword data and social discussions.

“What does the High impact pill mean?” It’s a relevance flag the system applies to the prompts it thinks are most worth your attention based on the run’s analysis. You can also filter to only these via Filters → High impact. “Why can’t I click the Start tracking button?” Only users with the admin role can track prompts. If you’re a viewer, the button stays disabled and the tooltip reads “You need admin access to perform this action”. “I clicked Export CSV after filtering: why does the CSV include all prompts?” The CSV export always contains every suggestion prompt in the run, regardless of the filter or search state on the page. If you need a filtered export, you can hide unwanted rows in your spreadsheet tool afterward. “The Added meter says 6/100: does that mean I can only add 100 prompts?” No. The denominator (100) is a fixed reference scale to keep the visual bar meaningful; it’s not a limit. You can add as many prompts as you like (subject to your plan’s prompt limits in your tracked library). “Where did this analysis come from?” It was generated by a Prompt Universe run: kicked off from the Prompt Universe tab on the Prompts page. The run scans your domain, your competitors, and relevant social discussions to surface prompts where AI models may be mentioning your competitors but not you. “Can I rerun an analysis to refresh it?” Not from this page: runs are one-time. Start a new run from the Prompt Universe tab on the Prompts page. The run selector here shows your 4 most recent runs.